FEDERAL TRADE COMMISSION v. THOMAS JEFFERSON UNIVERSITY
Opinion
IN THE UNITED STATES DISTRICT COURT FOR THE EASTERN DISTRICT OF PENNSYLVANIA
FEDERAL TRADE COMMISSION et al., Plaintiffs,
CIVIL ACTION v. NO. 20-01113
THOMAS JEFFERSON UNIVERSITY et al., Defendants.
PAPPERT, J. December 8, 2020
MEMORANDUM The Federal Trade Commission and Pennsylvania Office of Attorney General, collectively the Government, seek to preliminarily enjoin a proposed merger between Thomas Jefferson University and the Albert Einstein Healthcare Network pending an administrative determination of whether the combination violates Section 7 of the Clayton Act. The parties conducted extensive discovery and the Court held six days of evidentiary hearings which included the testimony of twenty witnesses and the presentation of voluminous documentary evidence. The Court also received from the parties and reviewed additional documents, declarations, deposition transcripts and other materials. Following the hearings, the parties submitted proposed findings of fact and conclusions of law and the Court allowed the parties several hours of oral argument. To obtain the relief it seeks, the Government must define a relevant geographic market—that area where potential buyers look for the goods or services they want— within which the likely competitive effects of the merger can be evaluated. That market’s definition is dependent on the special characteristics of the industry involved and the Court is required to take a pragmatic and factual approach in determining whether the Government has done it correctly. Of greatest importance to this case, the
market’s geographic scope must “correspond to the commercial realities of the industry at issue.” The healthcare industry’s market is represented by a “two-stage model of competition.” In the first stage, hospitals compete to be included in an insurer’s hospital network. In the second, hospitals compete to attract individual members of the insurers’ plans. This means that insurers, not patients seeking and receiving medical care, are the payors—those who will most directly feel the impact of the increased price of care. This is what the Third Circuit Court of Appeals has called the “commercial reality” of the uniquely structured healthcare industry. Patients are not irrelevant to a hospital system merger analysis; their choices and behavior can affect the bargaining leverage
that hospitals and insurers possess when they negotiate hospitals’ inclusion in insurers’ networks and the reimbursement rates insurers agree to pay hospitals. But as the entities bearing the immediate impact of the cost of medical care, the insurers’ perspective is extremely important in deciding whether a merger will substantially lessen the competition for healthcare in a proposed geographic market. The propriety of a relevant geographic market in this industry must therefore be assessed “through the lens of the insurers.” To establish its prima facie case, the Government must put forth enough evidence to prove that the insurers would not avoid a price increase in any one of the Government’s proposed markets by looking to hospitals outside those markets. The Government has not met this burden. It contends that a combination of its expert’s econometric algorithm and testimony primarily from two (of the region’s four) major commercial insurers shows that its geographic markets correspond to the
commercial realities of southeastern Pennsylvania’s competitive healthcare industry. But the expert’s calculations alone do not do so, and the insurers’ testimony is neither unanimous, unequivocal nor supported by the record as a whole. Their conclusory assertions that they would have to succumb to a price increase for services in the Government’s proposed markets instead of looking to healthcare providers outside those markets are not credible. The Court denies the Government’s request for a preliminary injunction. I A On September 14, 2018, Jefferson and Einstein signed a System Integration
Agreement, (JX0078), pursuant to which Jefferson will become Einstein’s sole member and ultimate parent. (Pls.’ Proposed Findings of Fact (“FF”) ¶ 3); (Defs.’ FF ¶ 5.) On February 27, 2020, the FTC initiated an administrative proceeding seeking to permanently enjoin the proposed merger. A merits trial in that action is presently scheduled to begin on March 8, 2021. See Order Granting Continuance, In re Thomas Jefferson University et al., File No. 181 0128, Dkt. No. 9392 (FTC Nov. 6, 2020). Seeking to pause the merger and preserve the status quo pending the administrative proceeding’s outcome, the Government filed this lawsuit requesting a preliminary injunction under Section 13(b) of the FTC Act, 15 U.S.C. § 53(b), and Section 16 of the Clayton Act, 15 U.S.C. § 26. (Compl. at 1–2, ECF No. 7.) The System Integration Agreement expires on the later of December 31, 2021 or, in the event of an appeal from this decision, sixty days after a final decision by the Court of Appeals. (Pls.’ FF ¶ 3); (JX0078-045.)
B Jefferson and Einstein operate in a densely populated, major metropolitan region. There are abundant healthcare options in southeastern Pennsylvania, including fifty-one hospitals dedicated to general acute care (“GAC”), children’s specialty care, orthopedics and cancer care. (Capps Rpt. App’x G.1 ¶ 544.) Philadelphia’s healthcare market is less consolidated than others around the country. See (Sept. 14, 2020 Hr’g Tr. (Markowitz (Reg’l Director Operations and Mktg., Cigna)) 65:7–24, ECF. No. 250). In 2018, Jefferson and Einstein were just two of thirteen health systems providing inpatient GAC services in the region.1 (Capps Rpt. App’x G Fig. 41.)
i Jefferson includes a nonprofit health system operating fourteen hospitals with 2,885 licensed beds in Pennsylvania and New Jersey. (Capps Rpt. ¶ 105.) Jefferson hospitals providing inpatient GAC services include its flagship, Thomas Jefferson University Hospital (“TJUH”) in Philadelphia and Abington Hospital and Abington- Lansdale Hospital in Montgomery County.2 (Defs.’ FF ¶ 1.) Jefferson provides
1 In 2019, American Academic Health System stopped operating after closing Hahnemann University Hospital and selling St. Christopher’s Hospital for Children to Drexel University and Tower Health. (Capps Rpt. App’x G Fig. 41.)
2 Jefferson’s system also includes Bucks Hospital, Cherry Hill Hospital (NJ), Frankford Hospital, Jefferson Hospital for Neuroscience, Methodist Hospital, Stratford Hospital (NJ), inpatient rehabilitation services in a twenty-three-bed unit at Abington Hospital and at the ninety-six-bed freestanding inpatient rehabilitation facility (“IRF”) Magee Rehabilitation Hospital, which is in Philadelphia. See (Pls.’ FF ¶ 1); (Defs.’ FF ¶ 1); (Ramanarayanan Rpt. ¶ 64, Ex. 3). Jefferson also operates urgent care centers,
outpatient centers, testing and imaging centers and a cancer center.3 (Capps Rpt. ¶ 105.) ii Einstein is a non-profit health system which includes three GAC hospitals: its 548-bed Einstein Medical Center Philadelphia (“EMCP”) in North Philadelphia, the sixty-seven-bed Einstein Medical Center Elkins Park (“EMCEP”) in southeastern Montgomery County and its 191-bed Einstein Medical Center Montgomery (“EMCM”) in East Norriton, Montgomery County. See (Capps Rpt. ¶¶ 114–128); (Pls.’ FF ¶ 2); (Defs.’ Answer to Compl. ¶ 35, ECF No. 51). EMCP accounts for seventy percent of Einstein’s revenues. See (Sept. 16, 2020
Hr’g Tr. (Freedman (CEO, Einstein)) 106:21–23, ECF No. 252). However, Einstein’s
Free access — add to your briefcase to read the full text and ask questions with AI
IN THE UNITED STATES DISTRICT COURT FOR THE EASTERN DISTRICT OF PENNSYLVANIA
FEDERAL TRADE COMMISSION et al., Plaintiffs,
CIVIL ACTION v. NO. 20-01113
THOMAS JEFFERSON UNIVERSITY et al., Defendants.
PAPPERT, J. December 8, 2020
MEMORANDUM The Federal Trade Commission and Pennsylvania Office of Attorney General, collectively the Government, seek to preliminarily enjoin a proposed merger between Thomas Jefferson University and the Albert Einstein Healthcare Network pending an administrative determination of whether the combination violates Section 7 of the Clayton Act. The parties conducted extensive discovery and the Court held six days of evidentiary hearings which included the testimony of twenty witnesses and the presentation of voluminous documentary evidence. The Court also received from the parties and reviewed additional documents, declarations, deposition transcripts and other materials. Following the hearings, the parties submitted proposed findings of fact and conclusions of law and the Court allowed the parties several hours of oral argument. To obtain the relief it seeks, the Government must define a relevant geographic market—that area where potential buyers look for the goods or services they want— within which the likely competitive effects of the merger can be evaluated. That market’s definition is dependent on the special characteristics of the industry involved and the Court is required to take a pragmatic and factual approach in determining whether the Government has done it correctly. Of greatest importance to this case, the
market’s geographic scope must “correspond to the commercial realities of the industry at issue.” The healthcare industry’s market is represented by a “two-stage model of competition.” In the first stage, hospitals compete to be included in an insurer’s hospital network. In the second, hospitals compete to attract individual members of the insurers’ plans. This means that insurers, not patients seeking and receiving medical care, are the payors—those who will most directly feel the impact of the increased price of care. This is what the Third Circuit Court of Appeals has called the “commercial reality” of the uniquely structured healthcare industry. Patients are not irrelevant to a hospital system merger analysis; their choices and behavior can affect the bargaining leverage
that hospitals and insurers possess when they negotiate hospitals’ inclusion in insurers’ networks and the reimbursement rates insurers agree to pay hospitals. But as the entities bearing the immediate impact of the cost of medical care, the insurers’ perspective is extremely important in deciding whether a merger will substantially lessen the competition for healthcare in a proposed geographic market. The propriety of a relevant geographic market in this industry must therefore be assessed “through the lens of the insurers.” To establish its prima facie case, the Government must put forth enough evidence to prove that the insurers would not avoid a price increase in any one of the Government’s proposed markets by looking to hospitals outside those markets. The Government has not met this burden. It contends that a combination of its expert’s econometric algorithm and testimony primarily from two (of the region’s four) major commercial insurers shows that its geographic markets correspond to the
commercial realities of southeastern Pennsylvania’s competitive healthcare industry. But the expert’s calculations alone do not do so, and the insurers’ testimony is neither unanimous, unequivocal nor supported by the record as a whole. Their conclusory assertions that they would have to succumb to a price increase for services in the Government’s proposed markets instead of looking to healthcare providers outside those markets are not credible. The Court denies the Government’s request for a preliminary injunction. I A On September 14, 2018, Jefferson and Einstein signed a System Integration
Agreement, (JX0078), pursuant to which Jefferson will become Einstein’s sole member and ultimate parent. (Pls.’ Proposed Findings of Fact (“FF”) ¶ 3); (Defs.’ FF ¶ 5.) On February 27, 2020, the FTC initiated an administrative proceeding seeking to permanently enjoin the proposed merger. A merits trial in that action is presently scheduled to begin on March 8, 2021. See Order Granting Continuance, In re Thomas Jefferson University et al., File No. 181 0128, Dkt. No. 9392 (FTC Nov. 6, 2020). Seeking to pause the merger and preserve the status quo pending the administrative proceeding’s outcome, the Government filed this lawsuit requesting a preliminary injunction under Section 13(b) of the FTC Act, 15 U.S.C. § 53(b), and Section 16 of the Clayton Act, 15 U.S.C. § 26. (Compl. at 1–2, ECF No. 7.) The System Integration Agreement expires on the later of December 31, 2021 or, in the event of an appeal from this decision, sixty days after a final decision by the Court of Appeals. (Pls.’ FF ¶ 3); (JX0078-045.)
B Jefferson and Einstein operate in a densely populated, major metropolitan region. There are abundant healthcare options in southeastern Pennsylvania, including fifty-one hospitals dedicated to general acute care (“GAC”), children’s specialty care, orthopedics and cancer care. (Capps Rpt. App’x G.1 ¶ 544.) Philadelphia’s healthcare market is less consolidated than others around the country. See (Sept. 14, 2020 Hr’g Tr. (Markowitz (Reg’l Director Operations and Mktg., Cigna)) 65:7–24, ECF. No. 250). In 2018, Jefferson and Einstein were just two of thirteen health systems providing inpatient GAC services in the region.1 (Capps Rpt. App’x G Fig. 41.)
i Jefferson includes a nonprofit health system operating fourteen hospitals with 2,885 licensed beds in Pennsylvania and New Jersey. (Capps Rpt. ¶ 105.) Jefferson hospitals providing inpatient GAC services include its flagship, Thomas Jefferson University Hospital (“TJUH”) in Philadelphia and Abington Hospital and Abington- Lansdale Hospital in Montgomery County.2 (Defs.’ FF ¶ 1.) Jefferson provides
1 In 2019, American Academic Health System stopped operating after closing Hahnemann University Hospital and selling St. Christopher’s Hospital for Children to Drexel University and Tower Health. (Capps Rpt. App’x G Fig. 41.)
2 Jefferson’s system also includes Bucks Hospital, Cherry Hill Hospital (NJ), Frankford Hospital, Jefferson Hospital for Neuroscience, Methodist Hospital, Stratford Hospital (NJ), inpatient rehabilitation services in a twenty-three-bed unit at Abington Hospital and at the ninety-six-bed freestanding inpatient rehabilitation facility (“IRF”) Magee Rehabilitation Hospital, which is in Philadelphia. See (Pls.’ FF ¶ 1); (Defs.’ FF ¶ 1); (Ramanarayanan Rpt. ¶ 64, Ex. 3). Jefferson also operates urgent care centers,
outpatient centers, testing and imaging centers and a cancer center.3 (Capps Rpt. ¶ 105.) ii Einstein is a non-profit health system which includes three GAC hospitals: its 548-bed Einstein Medical Center Philadelphia (“EMCP”) in North Philadelphia, the sixty-seven-bed Einstein Medical Center Elkins Park (“EMCEP”) in southeastern Montgomery County and its 191-bed Einstein Medical Center Montgomery (“EMCM”) in East Norriton, Montgomery County. See (Capps Rpt. ¶¶ 114–128); (Pls.’ FF ¶ 2); (Defs.’ Answer to Compl. ¶ 35, ECF No. 51). EMCP accounts for seventy percent of Einstein’s revenues. See (Sept. 16, 2020
Hr’g Tr. (Freedman (CEO, Einstein)) 106:21–23, ECF No. 252). However, Einstein’s
Torresdale Hospital and Washington Township Hospital (NJ). See Jefferson Health, We Are Jefferson, at 3 (Jan. 2020), https://hospitals.jefferson.edu/content/dam/health/PDFs/general/aboutus/ We-Are-Jefferson-1-08-20.pdf. (last visited Dec. 7, 2020).
3 Outpatient services are used by health systems to “feed inpatient services in the total continuum of care.” (JX0034, Buongiorno (EVP and CFO, Main Line Health) Dep. Tr. 159:11–12); see also (Sept. 29, 2020 Hr’g Tr. (Meyer (President Jefferson Health, Senior EVP Thomas Jefferson University (“TJU”)) 69:5–13 (explaining hospitals attract inpatients at outpatient locations “through the affiliation purchase or recruitment and employment of primary care doc[tor]s that they place into specific communities”)); (Capps Rpt. ¶ 63 (“[H]ospitals located outside of, but not overly far from, a given geography can attract patients from that area [through] local affiliated or owned medical groups . . . within the geography. These medical groups can act as ‘front doors’ that steer patients to the associated system’s hospital . . .”)); (Sep. 29, 2020 Hr’g Tr. (Klasko) 55:24–25 (“[M]ore, and more, and more, and more things are going to be moving to [an] outpatient environment.”)); (Id. at 22:15– 25 (“[T]he whole definition of what a hospital is . . . is changing rapidly.”)). commercially insured population is declining and many of EMCP’s commercially insured patients arrive through the hospital’s Emergency Department. (Defs.’ FF ¶ 30.) EMCP is viewed as a “safety net hospital” because it has one of the highest percentages of government-insured inpatients—eighty seven percent or more—among large
hospitals in the United States. (Defs.’ FF ¶ 30.) Among the more than 800 large GAC hospitals in the United States, only sixteen recently had a comparable percentage of government-insured patients and six of those were government-operated. (Capps Rpt. ¶ 87.) Medicare and medical assistance coverage “do not cover the cost” of patient care because government reimbursement rates do not keep up with Einstein’s inflationary costs. (Sept. 16, 2020 Hr’g Tr. (Freedman) 185:6–11.) Einstein concluded that it should seek a strategic partner in order to create scale to allow for savings that could improve its financial situation driven by its payor mix. See (id. at 115:17–116:6). Einstein also provides inpatient rehabilitation services through MossRehab at its EMCP and EMCEP locations. (Pls.’ FF ¶ 2); (Defs.’ FF ¶ 2.) MossRehab at Elkins
Park is a 130-bed freestanding IRF. See (Smith Rpt. ¶ 69); (Ramanarayanan Rpt. ¶ 64, Ex. 3.) MossRehab also has inpatient beds at Jefferson’s Frankford and Bucks Hospitals and at Doylestown Hospital. (Pls.’ FF ¶ 2); (Smith Rpt. ¶ 69); (Ramanarayanan Rpt. ¶ 64, Ex. 3.) iii Other area health systems include the University of Pennsylvania Health System–also known as Penn Medicine. In southeastern Pennsylvania, Penn Medicine includes six acute care hospitals and hundreds of outpatient facilities. (Capps Rpt. ¶ 130.) It also operates facilities in Lancaster County and New Jersey. (Id.) The 821- bed Hospital of the University of Pennsylvania (“HUP”) in Philadelphia’s University City is Penn Medicine’s largest hospital offering GAC services. (Id. at ¶ 131.) Penn Presbyterian Medical Center, with 331 staffed beds, also in University City, provides GAC services as well. (Id. at ¶ 133.) Pennsylvania Hospital, with 567 beds located in
Center City, is Penn Medicine’s second-largest Philadelphia hospital providing GAC services. (Id. at ¶ 134.) Penn Medicine is in the process of opening “The Pavilion,” an additional facility with more GAC beds, across the street from HUP. (Id. at ¶ 132.) The new facility is expected to give Penn Medicine a 250-bed net inpatient gain. (Id.) The health system also provides GAC services at Chester County Hospital, a 275-bed suburban community hospital, and recently replaced an existing outpatient location in Delaware County at Penn Medicine Radnor with a new facility from which outpatients requiring inpatient care are anticipated to turn to Penn Medicine’s Philadelphia hospitals. (Id. at ¶¶ 136– 37); (JX0065, Gustave (SVP Bus. Dev., Penn Medicine) Dep. Tr. 73:18–74:5.) In
addition, Good Shepherd Penn Partners manages a fifty-eight-bed inpatient rehabilitation unit, the Penn Institute for Rehabilitation Medicine, which is licensed through HUP in Philadelphia. (Ramanarayanan Rpt. ¶ 64, Ex. 3.) Main Line Health is a nonprofit health system with four hospitals offering GAC services: 370-bed Lankenau Medical Center in Wynnewood, Montgomery County, its largest; 287-bed Bryn Mawr Hospital, also in Montgomery County; 231-bed Paoli Hospital, in eastern Chester County; and 204-bed Riddle Hospital, to Philadelphia’s southwest in Delaware County. (Capps Rpt. ¶¶ 138–142); (Smith Rpt. ¶ 60.) Main Line Health also operates a 148-bed freestanding IRF at Bryn Mawr Rehabilitation Hospital (“Bryn Mawr Rehab”) in Chester County. See (Ramanarayanan Rpt. ¶ 64, Ex. 3). Tower Health operates six GAC hospitals in southeastern Pennsylvania: 148-bed Chestnut Hill Hospital in Philadelphia County; 232-bed Pottstown Hospital in western
Montgomery County; 139-bed Phoenixville Hospital in northern Chester County, just across the Montgomery County border; 714-bed Reading Hospital in Berks County; and 171-bed Brandywine Hospital and sixty-three-bed Jennersville Hospital in western Chester County. (Capps Rpt. ¶¶ 145–151); (Smith Rpt. ¶ 60.) Phoenixville Hospital also has fourteen inpatient rehabilitation beds. See (Ramanarayanan Rpt. ¶ 64, Ex. 3). Tower also operates twenty-two urgent care locations, including two in Montgomery County in Plymouth Meeting and Conshohocken. (JX0027, Ahern (EVP Business Development & Strategy, Tower Health) Dep. Tr. 16:4–5, 140:4–7.) Tower owns St. Christopher’s Hospital for Children, a 188-bed hospital in North Philadelphia that provides GAC services to children. (Smith Rpt. ¶ 60.)
Temple Health, a subsidiary of Temple University, is a nonprofit health system. (Capps Rpt. ¶ 165.) Its hospitals include Temple University Hospital, Temple University Hospital – Jeanes Campus (“Jeanes”), Fox Chase Cancer Center, Temple University Hospital – Episcopal Campus and Temple University Hospital – Northeastern Campus. (Id.) Temple University Hospital, the system’s largest with 732 beds, is in North Philadelphia. (Id. at ¶ 166); (Smith Rpt. ¶ 60.) Jeanes is a 146-bed hospital in northeast Philadelphia. (Capps Rpt. ¶ 167); (Smith Rpt. ¶ 60.) The Episcopal Campus, in Philadelphia’s Kensington neighborhood, is largely a behavioral health facility, but has an emergency room and offers other medical services. (Id. at ¶ 168.) Temple’s Northeastern Campus was formerly an impatient hospital, but now serves as an outpatient facility. (Id.) Trinity Health Mid-Atlantic also runs several GAC hospitals in the region. St. Mary Medical Center in Langhorne, Bucks County, with 373 beds, is its largest area
hospital. (Id. at ¶ 157.) Nazareth Hospital, with 231 beds, is in northeast Philadelphia. (Id. at ¶ 156.) It has twenty inpatient rehabilitation beds. (Ramanarayanan Rpt. ¶ 64, Ex. 3.) Mercy Fitzgerald, with 183 beds, is in Delaware County. (Capps Rpt. ¶ 155.) It has ten inpatient rehabilitation beds. See (Ramanarayanan Rpt. ¶ 64, Ex. 3). Trinity has also operated 157-bed Mercy Philadelphia Hospital in southwest Philadelphia, although it is slated to stop offering services there, with some services shifting to other area providers and others moving to Mercy Fitzgerald. (Capps Rpt. ¶ 154.) In 2018, the Mercy Health System and St. Mary formed a clinical affiliation with Penn Medicine, facilitating access to Penn Medicine services when required by their patients. (Id. at ¶ 159); (JX0065, Gustave Dep. Tr. 53:18–54:16.) Trinity also operates a fifty-bed
IRF at St. Mary Rehabilitation Hospital (“St. Mary Rehab”) in Bucks County. (Ramanarayanan Rpt. ¶ 64, Ex. 3.) Grand View Health operates Grand View Hospital, a 169-bed GAC hospital in Bucks County. (Capps Rpt. ¶ 173.) Grand View has fourteen inpatient rehabilitation beds. See (Ramanarayanan Rpt. ¶ 64, Ex. 3). In 2019, Grand View announced a $210 million, 170,000 square foot expansion to include a new emergency department, intensive care unit beds and private inpatient rooms. (Capps Rpt. ¶ 174.) Grand View, like St. Mary and Mercy Health, has also entered into a joint clinical partnership with Penn Medicine, part of Penn’s effort to “attract tertiary volume to come down to the Penn hospitals.” (JX0065, Gustave Dep. Tr. 53:14–21, 54:13–16.) The Prime Healthcare Foundation runs the nonprofit Suburban Community Hospital, a 126-bed hospital in Montgomery County. (Capps Rpt. ¶ 175.) Prime Healthcare also operates two for-profit hospitals in the area, Roxborough Memorial
Hospital in Philadelphia and Lower Bucks Hospital in Bucks County. (Id.) Holy Redeemer Health System, a nonprofit health system, operates one GAC hospital, Holy Redeemer, with 242 beds in Montgomery County along with ambulatory care sites in Bucks County, eastern Montgomery County and northeast Philadelphia. (Id. at ¶ 160.) Doylestown Hospital, with 232 beds in Bucks County, is the sole hospital in Doylestown Health’s nonprofit system. (Id. at ¶ 162.) It recently underwent a $100 million expansion, adding beds and other services. (Id. at ¶ 164.) Its oncology programs are part of a clinical partnership with Penn Medicine. (Id.) Cancer Treatment Centers of America, Philadelphia operates a twenty-two-bed
cancer hospital in North Philadelphia. (Smith Rpt. ¶ 60.) In addition to the hospital-affiliated IRFs identified above, the Kessler Institute for Rehabilitation (“Kessler Marlton”) operates a sixty-one-bed freestanding IRF in New Jersey. See (Ramanarayanan Rpt. ¶ 64, Ex. 3). Rehabilitation services are also offered in the area at a number of skilled nursing facilities (“SNFs”) that are not directly tied to the region’s health systems. Area SNF operators include Genesis Healthcare. See (id. at ¶ 103). Genesis operates thirty-eight SNFs in Pennsylvania, four of which are PowerBack Rehabilitation facilities, “designed to provide short-stay skilled nursing . . . to deliver a comprehensive rehabilitation regimen in accommodations specifically designed to serve high-acuity patients.” See (id.). The Pennsylvania PowerBack facilities are in Center City Philadelphia (PowerBack-Lombard, 150 beds), Montgomery County (PowerBack-Hatboro, 109 beds) and Chester County (PowerBack- Phoenixville, twenty-two beds and PowerBack-Exton, 120 beds). (See id. at ¶ 103
n.240). Rehab at Shannondell operates a 120-bed SNF on the campus of a senior assisted living community in Montgomery County. See (id. at ¶ 103). Also, Abramson Senior Care offers senior short-term rehabilitation services at the Abramson Residence and the Birnhak Transitional Care center at Lankenau, both in Montgomery County. See (id.). C The region’s commercial health insurance market is far more consolidated than the provider market. Jefferson’s Chief Executive Officer Dr. Stephen Klasko characterized the area as having “the worst externalities of any city in the country” for healthcare systems because there is “pretty much a monopolistic type insurance
situation with a few insurers.” See (Sept. 29, 2020 Hr’g Tr. (Klasko (CEO, Jefferson)) 19:21–20:2, ECF No. 261). The region has only four major commercial health insurance providers: Independence Blue Cross (“IBC”), Aetna, Cigna and United Healthcare (“United”). See (Capps Rpt. ¶¶ 177–191). Because healthcare provider competition in the area is extensive, Klasko explained that commercial insurers “especially the big ones, United, Aetna, IBC, of course, and Cigna, they could just say fine, we won’t [keep a provider in-network]” and not suffer negative repercussions. (Sept. 29, 2020 Hr’g Tr. (Klasko) 27:3–7.) IBC is the area’s dominant commercial insurer, with more than fifty percent market share covering approximately 1.3 million lives and coverage agreements with every area health system. (Defs.’ FF ¶ 74); see also (Sept. 14, 2020 Hr’g Tr. (Staudenmeier (VP Provider Contracting, IBC)) 62:14–21); (id. (Markowitz) at 61:11– 13); (PX5008, Staudenmeier Decl. ¶ 3); (JX0064, Winings (VP Network Management,
United) Dep. Tr. 281:12–15); (JX0062, Morris (VP Provider Networks, Aetna) Dep. Tr. 115:16–116:19, 124:10–17); (DX0127-002); (DX0405-003); (DX0317-10). At the evidentiary hearing, IBC could not identify a single health system that has been out of its coverage network for longer than six months. See (Sept. 14, 2020 Hr’g Tr. (Staudenmeier) 110:25–111:2). IBC has “a very strong market position” because there are significantly more other hospital options than other insurance options. (Sept. 14, 2020 Hr’g Tr. (DeAngelis (CFO, Jefferson)) 307:10–25.) All other major commercial insurers in southeastern Pennsylvania recognize IBC as the prevailing player in the commercial insurance market. See (Sept. 14, 2020 Hr’g Tr. (Markowitz) 61:14–18); (JX0062, Morris Dep. Tr. 124:15–17); (JX0064, Winings Dep. Tr. 142:7–12).
According to Aetna and United, healthcare providers fear IBC will retaliate against them if they partner with other payors by reducing benefits or terminating its relationships with them. See (JX0062, Morris Dep. Tr. 113:5–22 (explaining Jefferson and Penn had expressed concerns “about IBC retaliating” if they made certain coverage arrangements “with Aetna or any other carrier”)); (DX0442-003 (“IBC dominant player and all health systems have a ‘fear’ of Blue retribution if they were to align themselves in any way with another pay[o]r based on history.”)); (JX0064, Winings Dep. Tr. 283:4– 16 (explaining health systems had concerns about partnering with United out of fear that “IBC would either terminate them from the network” or make “meaningful and impactful” rate reductions)). Multiple witnesses testified that neither Jefferson nor Einstein can afford being out of IBC’s network. (Defs.’ FF ¶ 94.) At Jefferson, payments from IBC comprise approximately fifty-eight percent of commercial GAC revenues,4 roughly fifty percent of
its total commercial insurance reimbursements and approximately twenty percent of its total revenue. See (Capps Rpt. Fig. 2); (Sept. 14, 2020 Hr’g Tr. (DeAngelis) 289:6–14). An IBC short-term financial analysis showed that if Jefferson were not included in IBC’s network, the resulting harm to Jefferson could amount to tens of millions of dollars. See (Sept. 14, 2020 Hr’g Tr. (Staudenmeier) 111:16–112:6). It determined that cutting Jefferson out of its network would not impact its network adequacy from a regulatory standpoint. (Id. at 108:24–109:3.) IBC accounts for approximately fifty-seven percent of Einstein’s commercial GAC revenues and approximately nineteen percent of the system’s hospital revenues. See (Capps Rpt. Fig. 2); (Sept. 30, 2020 Hr’g Tr. (McTiernan (SVP Clinical and Provider
Management, Health Partners Plans, formerly at Einstein and IBC)) 60:22–23, ECF No. 262). An IBC analysis contemplating Einstein’s termination from its network showed that Einstein would lose tens of millions of dollars from termination and IBC would have sufficient network access and adequacy from a regulatory standpoint without Einstein. See (DX0329-008, -010). Aetna covers approximately 550,000 to 650,000 lives in the Philadelphia area.
4 Percent estimates of IBC’s contribution to Jefferson and Einstein commercial GAC revenues includes “the pay[o]r designation for plans associated with the Blue Cross and Blue Shield Association that are not classified as IBC.” (Capps Rpt. ¶ 177 n.206.) According to Dr. Capps, IBC itself accounts for forty percent of Jefferson’s commercial GAC revenue and forty-four percent of Einstein’s commercial GAC revenue. See (id. at Fig. 2). See (Capps Rpt. ¶ 184). It is the second largest commercial payor for both Jefferson and Einstein. See (id.). Aetna accounts for approximately twenty-five percent and twenty- nine percent of Jefferson and Einstein’s commercial GAC revenues, respectively. See (id. at Fig. 2). Its reimbursement payments constitute eight to ten percent of
Jefferson’s total revenue and approximately seven percent of Einstein’s hospital revenues. See (Sept. 14, 2020 Hr’g Tr. (DeAngelis) 289:15-22); (Sept. 30, 2020 Hr’g Tr. (McTiernan) 60:18–19). United is even smaller, covering approximately 300,000 lives in Philadelphia and Montgomery counties. See (PX5007, Winings Decl. ¶ 2). It accounts for roughly six percent of both Jefferson and Einstein’s commercial GAC revenues, three to four percent of Jefferson’s total revenue and one to two percent of Einstein’s hospital revenue. See (Capps Rpt. Fig. 2); (Sept. 14, 2020 Hr’g Tr. (DeAngelis) 289:23-290:1); (Sept. 30, 2020 Hr’g Tr. (McTiernan) 60:19–20). United excludes Jefferson from some of its commercial products and has been able to successfully market them. (Defs.’ FF
¶ 86.) It considered terminating its contracts with Einstein in early 2020 and determined that, for most of its plans, it could do so without creating patient access issues. (Id. at ¶ 87.) Of the four primary commercial insurers in southeastern Pennsylvania, Cigna is the smallest, covering approximately 200,000 lives and six percent of the commercial healthcare market in the five-county Philadelphia area. See (PX5006, Markowitz Decl. ¶ 2). Cigna accounts for approximately five percent of both Jefferson and Einstein’s commercial GAC revenues, one and a half to two percent of Jefferson’s total revenue and less than one percent of Einstein’s hospital revenue. See (Capps Rpt. Fig. 2); (Sept. 14, 2020 Hr’g Tr. (DeAngelis) 289:23-290:1); (Sept. 30, 2020 Hr’g Tr. (McTiernan) 58:12–15). D The Government proposes three relevant markets in which to assess the
proposed merger’s competitive effects. Two of the proposed markets are for inpatient GAC services sold to commercial insurers and their members and the third is for inpatient acute rehabilitation services sold to commercial insurers and their members. (Pls.’ FF ¶ 14.) Each proposed product market has different geographic boundaries. i GAC services include a broad cluster of medical, surgical, and diagnostic services that require an overnight hospital stay. (Pls.’ FF ¶ 15.) The parties agree that GAC services is a relevant product market. See (id. at ¶ 16); (Oct. 26, 2020 Oral Arg. Tr. 257:9–10, ECF No. 273). Insurers include local GAC hospitals in their networks because patients prefer to receive GAC services near their homes.5 (Pls.’ FF ¶ 21.)
The FTC does something in this case that it has never attempted in an effort to block a merger in the healthcare industry—allege multiple geographic markets for the same product, here GAC services. See (Oct. 26, 2020 Oral Arg. Tr. at 180–81). The Government includes three of the same hospitals in overlapping markets, magnifying their competitive significance. See (Smith Rpt. Figs. 3 and 4); (Capps Rpt. ¶¶ 35–36);
5 While insurers agreed that patients prefer to seek care “close to home,” when asked to define the boundaries of “close to home” care, insurers could not do so. See (JX0064, Winings Dep. Tr. 72:14–75:16 (“close to home” depends on access standards required by the Department of Health or other entities and how far patients are willing to travel, but United has not studied how far patients are willing to travel for GAC services)); (JX0070, Staudenmeier Dep. Tr. 175:5–9 (no objective standard in mind for statement that patients “generally prefer to receive care close to home for most routine inpatient and outpatient services”)). (id. at App’x G.1 ¶ 544 (noting Abington, Chestnut Hill and Roxborough Memorial are included in both alleged GAC markets)). a The Government first attempts to define what it terms the “Northern
Philadelphia Area” market, in which it includes eleven hospitals: Einstein’s EMCP and EMCEP; Jefferson’s Abington and Frankford Hospitals; Prime’s Roxborough Memorial Hospital; Temple University Hospital; Jeanes; Tower Health’s Chestnut Hill Hospital; Fox Chase Cancer Center; Cancer Treatment Centers of America, Philadelphia; and St. Christopher’s Hospital for Children. (Pls.’ FF ¶ 23.) Notably, Abington sits on the edge of the market at its far northern end. (Capps Rpt. ¶ 35.) This market does not include, among others, Penn Medicine’s three Philadelphia hospitals, notwithstanding that the area from which the hospitals draw seventy-five percent of their patients—the Patient Service Area (“PSA”)—for all three of Penn Medicine’s hospitals includes EMCP. (Defs.’ FF ¶ 27); see also (Capps Rpt. ¶ 37 (defining “patient service areas”)). It also excludes
Holy Redeemer Hospital, (Defs.’ FF ¶¶ 26, 28), even though its PSA encompasses North Philadelphia. (Defs.’ FF ¶ 28.) b The Government’s proposed “Montgomery Area” market for GAC services also includes Jefferson’s Abington Hospital, Prime’s Roxborough Memorial Hospital and Tower Health’s Chestnut Hill Hospital along with seven other hospitals: Jefferson’s Abington Lansdale Hospital; Einstein’s EMCM; Main Line Health’s Bryn Mawr and Paoli Hospitals; Prime’s Suburban Community Hospital; Tower Health’s Phoenixville Hospital; and Physician’s Care Surgical Hospital. (Pls.’ FF ¶ 26.) Abington sits on the edge of this market as well, this time at its far eastern end. (Capps Rpt. ¶ 35.) The market does not include, among others, Lankenau Hospital, Pottstown Hospital, Grand View, Doylestown Hospital, Jeanes or any Penn Medicine facility. (Defs.’ FF ¶¶ 14, 23– 24); (Capps Rpt. Fig. 42.)
c While Einstein aspires to compete with Jefferson, (PX2146-011), Jefferson identifies its primary competition as Penn Medicine, Main Line Health, Temple University and Tower Health. See (Sept. 29, 2020 Hr’g Tr. (Meyer) 63:7–11). It does not consider Einstein to be “a primary competitor for commercial patients because their commercial pay[o]r mix is so small. And their commercial payer mix comes almost entirely from their emergency room . . . . we don’t compete with them for elective cases because less than 1 percent of their volume is actually that kind of elective commercial case.” (Id. at 63:17–64:5.) EMCM is not a primary competitor for Jefferson’s Abington Hospital because I-476 acts as a dividing line for where patients seek care—Abington is
east of I-476 and EMCM is west of I-476. See (id. at 65:13–66:6); see also (Sept. 30, 2020 Hr’g Tr. (Merlis (EVP Strategic Partnerships, Strategic Ventures and Innovation, Jefferson) 119:9–15); (Capps Rpt. ¶¶ 37–38, 41–42). Jefferson sees Abington’s primary competitors as Grand View Hospital and Doylestown Hospital, Holy Redeemer Hospital, “maybe to a much smaller extent Chestnut Hill [Hospital], and to a smaller extent Main Line Health.” See (Sept. 29, 2020 Hr’g Tr. (Meyer) 64:23–65:2). In Jefferson’s view, Abington-Lansdale’s primary competitors are Grand View and Doylestown. See (id. at 66:14–16). Most insurers recognize Penn Medicine as Jefferson’s closest competitor. (Defs.’ FF ¶ 90); see also (Sept. 14, 2020 Hr’g Tr. (Markowitz) 67:13–15); (JX0062, Morris Dep. Tr. 107:18–108:4 (Jefferson and Penn “believe that their main competitor is each other”)). During negotiations over coverage agreements, payors leverage Jefferson and Penn against each other. They compare Jefferson to Penn or Temple, and not to
Einstein. See (Defs.’ FF ¶ 90); (JX0064, Winings Dep. Tr. 225:12–16). ii The Government’s third proposed relevant market is for inpatient acute rehabilitation services (“Acute Rehabilitation Services”) sold and provided to commercial insurers and their members in what it terms the “Philadelphia Area.” (Pls.’ FF ¶ 35). This represents another first for the FTC—it has never before litigated a case where it has attempted to define rehabilitation services as a relevant product market. See (Oct. 26, 2020 Oral Arg. Tr. 75:17–21); see also (Defs.’ FF ¶ 55); (Sept. 29, 2020 Hr’g Tr. (Ramanarayanan) 249:21–22). The Acute Rehabilitation Services product market includes only inpatient
rehabilitation services provided at IRFs: a “cluster of intensive inpatient rehabilitation therapy services that include, at a minimum, multi-disciplinary therapy at least three hours a day for five days per week, three face-to-face visits with a physician per week, and 24-hour nursing care.” (Pls.’ FF ¶ 36); (Smith Rpt. ¶¶ 118–120.) The product market excludes inpatient rehabilitation services provided at SNFs, which the Government defines as “non-hospital post-acute care settings that provide short-term and long-term nursing services and, at some SNFs, subacute rehabilitation services.” (Pls.’ FF ¶ 37.) The product market also excludes what Defendants call “high-end” SNFs, or “SNFs that have the capabilities to treat medically complex patients with many of the same conditions as those admitted by IRFs and provide inpatient rehabilitation services similar in nature and intensity (i.e., hours of therapy per day) to those offered at IRFs.” (Ramanarayanan Rpt. ¶ 17.B.i.) The Government claims that “Acute Rehab Services are provided only at IRFs.”
(Pls.’ FF ¶ 37.) According to the Government, IRF Acute Rehabilitation Services are “[d]emanded by [d]istinct [c]ustomers,” have “[d]istinct [c]haracteristics” compared to SNF services, are recognized “as a [d]istinct [l]evel of [p]ost-[a]cute [c]are” by industry participants, are provided at “[u]nique [f]acilities by [s]pecialized [v]endors,” and have distinct prices relative to other services. See (id. at ¶¶ 38–49). a The proposed Philadelphia Area geographic market for Acute Rehabilitation Services includes seven IRFs: three freestanding—Magee, MossRehab at Elkins Park and “Penn Rehab” (also known as Good Shepherd Penn Partners)—and five hospital- based—Jefferson Frankford (MossRehab), EMCP (MossRehab), Abington and Trinity’s
Nazareth Hospital. See (Smith Rpt. ¶ 147, Fig. 5); see also (Ramanarayanan Rpt. ¶ 64, Ex. 3). The Government’s proposed geographic market does not include freestanding IRFs at Bryn Mawr Rehab, St. Mary Rehab or Kessler Marlton. (Defs.’ FF ¶ 59.) b MossRehab views Bryn Mawr Rehab and Kessler Marlton as competitors. See (DX8652 (Einstein email identifying Kessler Marlton as MossRehab’s “second competitor behind [Bryn Mawr Rehab]”)). Also, “MossRehab captures [fewer] commercial patients and more Medicaid patients when compared to Magee and Bryn Mawr.” (DX8517-037.) MossRehab defines its service area to include all of Philadelphia, Bucks, Chester, Delaware and Montgomery Counties. See (Ramanarayanan Rpt. ¶ 122); (DX8517-030). MossRehab nurse liaisons compete for patients with nurse liaisons from Bryn Mawr Rehab—they are “always there”—and from St. Mary, Kessler-Marlton, Good Shepherd Penn Partners and HealthSouth,
which has facilities in Delaware. See (Sept. 29, 2020 Hr’g Tr. (Seminara) 222:1–8). “Oftentimes” the nurse liaisons are “all following the same patient” until they decide where to turn for acute inpatient rehabilitation after a hospital stay. (Id. at 222:10– 13.) Patients referred to Moss may end up at Magee or at St. Mary or Penn Partners. See (id. at 237:10–14). Magee identifies its service area as including “the five-county Philadelphia region, Gloucester, Camden and Burlington Counties in New Jersey and New Castle County, Delaware.” (DX9402-004.) Magee classifies MossRehab and Bryn Mawr Rehab as “medium” risk rehabilitation competitors, citing Bryn Mawr’s “rankings, physical environment [and] amenities” and MossRehab’s “rankings, marketing to consumers and
physicians” and its “Centers of Excellence” and outpatient facilities. See (DX9337-028). Magee describes itself as the “third largest” of five inpatient acute rehabilitation facilities in the region, comparing itself to MossRehab, Bryn Mawr Rehab, Good Shepherd Penn Partners and St. Mary Rehab. (DX9402-004.) Insurers acknowledge that their customers are not primarily interested in inpatient rehabilitation offerings when selecting health plans. See (Sept. 14, 2020 Hr’g Tr. (Staudenmeier) 97:3–11). Although patients may prefer acute rehabilitation services near their homes, (Pls.’ FF ¶ 52), for insurers what matters is offering a plan with in-network acute rehabilitation at any location. To operate as a health plan, insurers must meet state and federal requirements, which include having facilities for acute rehabilitation. See (Sept. 14, 2020 Hr’g Tr. (Staudenmeier) 90:18–91:7 (“[I]t’s a requirement.”)). No insurer testified that they limit their in-network acute rehabilitation offerings to MossRehab and Magee. See, e.g., (JX0066, Markowitz Dep.
Tr. 216:20–217:15 (acknowledging Cigna contracts with Bryn Mawr Rehab, St. Mary Rehab and Good Shepherd Penn Partners)). II Section 7 of the Clayton Act prevents mergers “where in any line of commerce or in any activity affecting commerce in any section of the country, the effect of such acquisition may be substantially to lessen competition or tend to create a monopoly.” 15 U.S.C. § 18. Section 7 is concerned with “probabilities, not certainties”; its definition of antitrust liability is “relatively expansive.” Brown Shoe Co. v. United States, 370 U.S. 294, 323 (1962); FTC v. Penn State Hershey Med. Ctr., 838 F.3d 327, 337 (3d Cir. 2016). To stop a merger, a Section 7 plaintiff ultimately must show that a “substantial
lessening of competition” is “sufficiently probable and imminent.” United States v. Marine Bancorporation, Inc., 418 U.S. 602, 622, 623 n.22 (1974) (internal quotation omitted). If the FTC “has reason to believe” that a “corporation is violating, or is about to violate,” Section 7, Section 13(b) of the FTC Act allows it to move for a preliminary injunction preventing a merger pending the FTC’s administrative adjudication of the merger’s legality, as it has done in this case.6 15 U.S.C. § 53(b). The FTC need not
6 The Office of Attorney General seeks injunctive relief for a threatened antitrust violation pursuant to Section 16 of the Clayton Act, 15 U.S.C. § 26. See (Pls.’ Proposed Conclusions of Law ¶ 6). Different standards for preliminary relief apply under Section 13(b) and Section 16. Section 16 requires the Attorney General to show that it is likely to suffer irreparable harm without relief. See establish that the “proposed merger would in fact violate [S]ection 7.” Penn State Hershey, 838 F.3d at 337 (citation and internal quotation omitted). Rather, a preliminary injunction may issue “[u]pon a proper showing that, weighing the equities and considering the Commission’s likelihood of ultimate success, such action would be
in the public interest.” 15 U.S.C. § 53(b). Section 13(b)’s purpose “is to preserve the status quo and allow the FTC to adjudicate the anticompetitive effects of the proposed merger in the first instance.” Penn State Hershey, 838 F.3d at 352. “[D]oubts are to be resolved against the transaction.” Id. at 337 (citation and internal quotation omitted). The Government has the prima facie burden to show it is likely to succeed on the merits of its claim that the merger is anticompetitive. See id. To meet its burden, it “must: (1) propose the proper relevant market and (2) show that the effect of the merger in that market is likely to be anticompetitive.” Id. at 337–38; see also FTC v. RAG- Stiftung, 436 F. Supp. 3d 278, 291 (D.D.C. 2020) (“[T]he FTC bears the prima facie burden of showing that [a] merger will lead to undue concentration in the market for a
particular product in a particular geographic area.”) (citation and internal quotation omitted). “Statistics reflecting the shares of the market controlled by the industry leaders and the parties to the merger are, of course, the primary index of market power; but only a further examination of the particular market—its structure, history and probable future—can provide the appropriate setting for judging the probable anticompetitive effect of the merger.” Brown Shoe, 370 U.S. at 322 n.38. If the FTC establishes a likelihood of success on the merits, it creates a presumption in favor of preliminary injunctive relief. See FTC v. H.J. Heinz Co., 246
Ferring Pharms., Inc. v. Watson Pharms., Inc., 765 F.3d 205, 210 (3d Cir. 2014). F.3d 708, 726 (D.C. Cir. 2001). “Defendants can . . . rebut this presumption by demonstrating that the FTC’s prima facie case and market-share statistics inaccurately predict the merger’s probable effects in the relevant market.” RAG-Stiftung, 436 F. Supp. 3d at 291. They must show “either that the combination would not have
anticompetitive effects or that the anticompetitive effects of the merger will be offset by extraordinary efficiencies resulting from the merger.” Penn State Hershey, 838 F.3d at 347. If Defendants rebut the prima facie case, the burden of production returns to the Government, joining with the burden of persuasion which the Government always has. See id. at 337 (citations omitted). Even if the FTC establishes a likelihood of success on the merits, the Court “‘must still weigh the equities in order to decide whether enjoining the merger would be in the public interest.’” Id. at 352 (quoting H.J. Heinz Co., 246 F.3d at 726); see also 15 U.S.C. § 53(b). The Court considers “whether the injunction, not the merger, would be in the public interest.” Penn State Hershey, 838
F.3d at 353 (emphasis in original). “The question is whether the harm that the Hospitals will suffer if the merger is delayed will, in turn, harm the public more than if the injunction is not issued.” Id. at 352. III To determine whether a proposed merger is reasonably likely to violate the Clayton Act, it is first necessary to determine the relevant geographic and product markets. See id. at 338 (citing Marine Bancorporation, 418 U.S. at 618). Market definition allows “measurement of market shares and market concentration,” which “is not an end in itself, but is useful to the extent it illuminates [a] merger’s likely competitive effects.” U.S. Dep’t of Justice & Fed. Trade Comm’n, Horizontal Merger Guidelines § 4 (rev. Aug. 19, 2010) (“Merger Guidelines”).7 Indeed, market definition is “critical” because a proposed merger’s legality “almost always depends upon the market power of the parties involved.” FTC v. Cardinal Health, Inc., 12 F. Supp. 2d 34, 45
(D.D.C. 1998). As prescribed by Congress, definition of relevant markets is a “pragmatic” and “factual” exercise and “not a formal, legalistic one.” Brown Shoe, 370 U.S. at 336. It is “‘dependent upon the special characteristics of the industry involved.’” Penn State Hershey, 838 F.3d at 335 (quoting St. Alphonsus Med. Ctr.-Nampa Inc. v. St. Luke's Health Sys., Ltd., 778 F.3d 775, 783 (9th Cir. 2015) (further citation omitted).) In other words, a properly identified relevant market must “correspond to the commercial realities of the industry.” Brown Shoe, 370 U.S. at 336 (internal quotations omitted); Penn State Hershey, 838 F.3d at 338. As the Merger Guidelines explain, “merger analysis does not consist of uniform application of a single methodology. Rather, it is a fact-specific process through which the Agencies, guided by their
extensive experience, apply a range of analytical tools to the reasonably available and reliable evidence to evaluate competitive concerns in a limited period of time.” Merger Guidelines § 1. Because prices are established through multiple layers of competition, price setting is complicated in the commercial reality of the healthcare market. See Penn
7 The Merger Guidelines are not binding but may be used as persuasive authority. See United States v. Anthem, Inc., 855 F.3d 345, 349 (D.C. Cir. 2017) (“[A]s the Justice Department acknowledges, the court is not bound by, and owes no particular deference to, the [Horizontal Merger] Guidelines . . .”); St. Alphonsus Med. Ctr.-Nampa, Inc. v. St. Luke’s Health Sys., Ltd., 778 F.3d 775, 784 n.9 (9th Cir. 2014) (“Although the Merger Guidelines are ‘not binding on the courts,’ . . . they ‘are often used as persuasive authority.’”) (citations omitted); ProMedica Health Sys. v. FTC, 749 F.3d 559, 565 (6th Cir. 2014) (finding the Horizontal Merger Guidelines to be “useful but not binding upon us here”). State Hershey, 838 F.3d at 342. “In the first stage, hospitals compete to be included in an insurance plan’s hospital network. In the second stage, hospitals compete to attract individual members of an insurer’s plan.” Penn State Hershey, 838 F.3d at 342; see also (Sept. 29, 2020 Hr’g Tr. (Capps) 108:20–109:1).
Economic analysis that reliably illuminates the likely competitive effects of a merger requires measuring shares that are aligned with industry characteristics. As Defendants’ expert economist Dr. Corey S. Capps explained, “in this case that means aligning with two-stage competition and keeping the focus on stage one [competition] with the insurers as the customers.” (Sept. 29, 2020 Hr’g Tr. (Capps) 123:11–21 (emphasis added).) The Government’s expert economist Dr. Loren K. Smith likewise explained that the appropriate focus “is the extent to which the merger will change the relative bargaining power of . . . the provider versus the insurer and how that will affect prices.” (Sept. 15, 2020 Hr’g Tr. (Smith) 75:11–14); see also (id. at 74:2–3 (“healthcare services are generally sold in two stages or competed for in two stages”)); (id. at 74:16–
22 (the customers in a healthcare provider merger are most directly the insurance companies)); (Pls.’ FF ¶ 6.) “[T]he vast majority of health care consumers are not direct purchasers of health care—the consumers purchase health insurance and the insurance companies negotiate directly with the providers . . . .” Saint Alphonsus Med. Ctr, 778 F.3d at 784 (internal quotations and citation omitted). “[I]insurance companies effectively act both as buyers and sellers.” Id. The insurers are the most relevant buyers, who “must consider both whether employers would offer their plans and whether employees would sign up for them.” FTC v. Advocate Health Care Network, 841 F.3d 460, 475 (7th Cir. 2016). Not until the insurer passes [a price] increase on to the patient in the form of higher premiums will the patient feel the impact of that price increase. And even then, the cost will be spread among many insured patients; it will not be felt solely by the patient who receives the higher-priced procedure. This is the commercial reality of the healthcare market as it exists today.”
Penn State Hershey, 838 F.3d at 342; see also Advocate Health Care, 841 F.3d at 471 (“Insured patients are usually not sensitive to retail hospital prices, while insurers respond to both prices and patient preferences.”). “Patients, of course, are relevant.” Penn State Hershey, 838 F.3d at 343. But the Court’s focus is properly “on the likely response of insurers” and not on the “likely response of patients to a price increase.” Id. at 339. Ultimately, the Government’s case turns on whether their proposed markets correspond to this commercial reality. A The Government must establish the relevant geographic market, defined as the “area in which a potential buyer may rationally look for the goods or services he seeks.” Id. at 338 (internal quotations omitted). The market must contain the sellers or producers who are able “to deprive each other of significant levels of business” and is where the merger’s effect “on competition will be direct and immediate.” Advocate Health Care, 841 F.3d at 468 (internal quotations and citations omitted). The hypothetical monopolist test (“HMT”) is a “common method” used to define the relevant geographic market for evaluating a plaintiff’s likelihood of ultimate success with respect to claims that a merger will substantially lessen competition. Penn State Hershey, 838 F.3d at 338; see also Merger Guidelines § 4. It is not “the only test that the district courts may use in determining whether the Government has met its burden to properly define the relevant geographic market.” Penn State Hershey, 838 F.3d at 345. Here, however, the parties agree it is appropriate for the Court to consider whether the Government’s expert properly established relevant markets using the HMT. See (Oct. 26, 2020 Oral Arg. Tr. 4:21–5:2, 14:21–23, 31:19–21, 38:14–39:13, 264:9–13).
The HMT “asks what would happen if a single firm became the only seller in a candidate geographic region.” Advocate Health Care, 841 F.3d at 468 (citation omitted). If that single firm–the hypothetical monopolist–could profitably raise prices above competitive levels, the candidate geographic region is a relevant geographic market. See id. “[I]f customers would defeat the attempted price increase by buying from outside the region, it is not a relevant market” and the process “should be repeated with ever-larger candidates until it identifies a relevant geographic market.” Id. (citations omitted). Using the HMT, a proposed market is properly defined if a hypothetical monopolist could impose a small but significant non-transitory increase in price (“SSNIP”), typically an increase of five percent or more, within its proposed boundaries.
See Penn State Hershey, 838 F.3d at 338; see also Merger Guidelines, §§ 4.1.2, 4.2.1. A proposed geographic market is too narrow if “consumers would respond to a SSNIP by purchasing the product from outside the market, thereby making the SSNIP unprofitable.” Penn State Hershey, 838 F.3d at 338. Importantly, the Merger Guidelines “do[ not] tell economists exactly how they should identify a candidate market before performing” the HMT. (Sept. 15, 2020 Hr’g Tr. (Smith) 89:4–11.) B The Government’s candidate GAC markets focus more on patients, not the insurers who will bear the immediate impact of any price increases. When Dr. Smith selected his candidate markets, he considered “how closely substitutable healthcare providers are to groups of patients” because, in his opinion, insurer demand to have particular healthcare providers in-network “is derived from patient demand for those providers.” (Smith Rpt. ¶ 103.) Relying on insurer declarations, Dr. Smith concluded
that “whether an insurer considers healthcare providers to be close substitutes derives from whether the insurer’s health plan members consider those providers to be close substitutes.” (Id. ¶ 84); see also (PX5006, Markowitz Decl. ¶ 9); (PX 5008, Staudenmeier Decl. ¶ 9); (PX5007, Winings Decl. ¶ 13). Dr. Smith posited that “[p]atient substitution patterns play a critical role in determining the extent to which commercial insurers can credibly threaten to exclude a provider from their networks during a negotiation over prices and other terms.” (Id. at ¶ 90.) He contends that “[i]n the context of two-stage competition among healthcare providers, market definition requires using patient substitution patterns to identify a collection of close substitute facilities that is just large enough that a hypothetical monopolist of that set of facilities
could profitably impose a SSNIP in negotiations with insurers.” (Id. at ¶ 102.) Accordingly, Dr. Smith used diversion ratios, which are “a measure of patient substitution patterns” to define the relevant geographic markets for GAC. See (id. at ¶¶ 137(i), 142). The Government contends that the commercial realities of the healthcare industry in southeastern Pennsylvania are “baked into the diversion numbers” such that the candidate markets in which Dr. Smith implemented his HMT have “already accounted for” those realities. (Oct. 26, 2020 Oral Arg. Tr. 215:24–25, 267:16–25.) But the Court’s geographic market determination is not merely a “statistical exercise” looking for a hypothetical monopolist that can impose a SSNIP. See (Oct. 26, 2020 Oral Arg. Tr. 264:3–8). Market definition can rest on a mathematical equation only if the variables used in the equation reflect the market’s commercial realities. Diversion ratios only capture insurer preferences for the purpose of constructing a relevant
geographic market where there is evidence to show that insurer decisions about which hospitals to include in their networks are aligned with patient decisions about where to seek care. Although diversion ratios are one piece of evidence, they do not completely capture the commercial realities of a healthcare market with two-stage competition and provider/insurer dynamics like those in southeastern Pennsylvania. See (Oct. 26, 2020 Oral Arg. Tr. 217:23–218:3). The Court must consider the Government’s application of the HMT to their proposed geographic markets for GAC “through the lens of the insurers . . . .” Penn State Hershey, 838 F. 3d at 342. And “measures of patient substitution like diversion ratios do not translate neatly into options for insurers.”
Advocate Health Care, 841 F. 3d at 475. The Government’s reliance on patient diversion ratios undermines its GAC geographic market definition from the very start of Dr. Smith’s four-step analytical framework. He began his first step with EMCP and then “add[ed] to the candidate market EMCP’s closest substitute Jefferson hospital measured by diversion ratio.” (Smith Rpt. ¶ 137(i) (emphasis added).) In step two, Dr. Smith “add[ed] all hospitals to the candidate geographic market that are closer substitutes to EMCP in terms of diversion ratio than the Jefferson hospital identified in step [one].” (Id. at ¶ 137(ii) (emphasis added).) At step three, he “add[ed] all hospitals to the candidate market that are located closer to EMCP in terms of drive distance than the farthest hospital included in steps one and two.” (Id. at ¶ 137(iii).) Finally in his fourth step, Dr. Smith purported to “test whether the candidate geographic market satisfie[d] the [HMT] through a price increase at EMCP—i.e., whether a hypothetical monopolist of all
hospitals in the candidate market could profitably increase prices at EMCP by a SSNIP of at least [five percent] in negotiations with insurers.” (Id. at ¶ 137(iv) (emphasis added).) If the candidate market did not satisfy the HMT, Dr. Smith would repeat steps two through four, starting over at step two to add “the next closest substitute hospital by diversion ratio that is not already included in the candidate market” until identifying “a set of hospitals that satisfie[d] the [HMT].” (Id. at ¶ 137(iv) (emphasis added).) Following Dr. Smith’s formula, EMCP’s closest Jefferson substitute becomes Abington Hospital, with a diversion ratio of 20.3 percent from EMCP to Abington. (Id. at ¶ 141.) From that starting point, Dr. Smith constructed the “Northern Philadelphia
Area,” an area surrounding EMCP as one of two relevant GAC geographic markets. See (id. at Fig. 3). He defined the boundaries of “the Montgomery Area,” the other relevant GAC geographic market, in the same way, starting with EMCM at the center. (Id. at ¶ 141, Fig. 4.) In Smith’s multiple market construction, however, the Montgomery Area overlaps with Northern Philadelphia because “Abington Hospital is also EMCM’s closest Jefferson substitute, with a diversion ratio of 15.7% from EMCM to Abington Hospital.”8 (Id. at ¶ 141.) As a result of his calculations, both of his candidate markets
8 According to Dr. Capps, Dr. Smith’s inclusion of Abington in both proposed GAC services markets significantly overstates its competitive significance. See (Capps Rpt. ¶ 264). satisfy the HMT.9 See (id. at ¶ 141). No one disputes that the geographic market boundaries which arise from Dr. Smith’s calculations result in SSNIP values that satisfy the HMT. See (Pls.’ FF ¶ 19); (Sept. 29, 2020 Hr’g Tr. (Capps) 166:18–167:8). That is all the Government believes it
has to show: that once any geographic market which can be drawn passes the HMT, it can move to the next step and calculate market share. See (Oct. 26, 2020 Oral Arg. Tr. 7:2–6). The Government acknowledges, however, that the geographic market which passes the HMT must correspond with commercial realities. See (id. at 13:1–24, 160:5– 8); (Pls.’ FF ¶ 17). And as Dr. Capps explained, “some markets can pass [the HMT] and be more logical with respect to competitive realities and others can be less [so].” (Sept. 29, 2020 Hr’g Tr. (Capps) 170:1–4.) Econometric evidence “can be powerful evidence, but it is not the only evidence that courts consider in defining the relevant market.” United States v. Aetna Inc., 240 F. Supp. 3d 1, 39 (D.D.C. 2017). Because Dr. Smith relies on patient diversion ratios to
construct his candidate markets in step one and two of his model, the results of his algorithm do not in and of themselves address, much less answer, the relevant antitrust question, which is whether a hypothetical monopolist could profitably impose a SSNIP without insurance companies turning to providers outside the geographic markets. Specifically, Dr. Smith does not show whether “enough insurers, in the face of
9 According to Dr. Smith’s analysis, “the aggregate diversion ratio from EMCP to” the eleven Northern Philadelphia hospitals is “more than double the [twenty percent] aggregate diversion ratio necessary for a hypothetical monopolist to profitably increase” EMCP’s prices by a five percent SSNIP. (Smith Rpt. ¶ 141.) “[T]he aggregate diversion ratio from EMCM” to the ten Montgomery Area hospitals is “approximately three times the [twenty percent] aggregate diversion ratio necessary for a hypothetical monopolist to profitably increase” EMCM’s prices by a five percent SSNIP. (Id.) a [SSNIP], would avoid the price increase by looking to hospitals outside the proposed geographic market . . . .” Penn State Hershey, 838 F.3d at 342. Courts must consider whether insurer and patient behavior is “correlated.” Id. at 343. The Merger Guidelines instruct the Agencies that “[w]hen direct customers of
the merging firms,”—here insurers—“compete against one another in a downstream market, their interests may not be aligned with the interests of final consumers,”—here patients—“especially if the direct customers expect to pass on any anticompetitive price increase.” Merger Guidelines § 2.2.2. There is a “fundamental difference between analyzing the likely response of consumers through the patient or payor perspective.” Penn State Hershey, 838 F.3d at 342. Again, insurers are the ones who negotiate directly with the hospitals to determine both reimbursement rates and the hospitals that will be included in their networks.” Id. “[I]t is possible” that patient and payor behavior align “in some healthcare market[s],” but a conclusion that there is “correlated behavior” between patients and payors must be based on specific evidence before using
diversion ratios to construct a relevant market. Id. at 343. In Advocate Health Care, the Seventh Circuit held that the district court incorrectly assumed that diversion ratios and insurer options were correlated. 841 F.3d at 475. The Circuit also cautioned the defendant hospitals that they “overlook[ed] insurers’ role in the marketplace” when, on appeal, the hospitals used diversion ratios to support their argument that the government should have included Northwestern Memorial hospital in its candidate market as “the closest substitute for some NorthShore hospitals.” Id. at 475 n.5. The court explained that even if it took “the diversion ratios to mean that a sizable minority of patients consider Northwestern Memorial as a close substitute, it [did] not follow that insurers could offer it as a sufficient substitute for a commercially viable insurance network.” Id. On remand, the district court found that the government’s expert could rely on diversion ratios without “fatally undermin[ing]” his geographic market analysis even
though they do not always “translate neatly into options for insurers.” FTC v. Advocate Health Care, No. 15-11473, 2017 WL 1022015, at *4 (N.D. Ill. Mar. 16, 2017). There, however, the court was able to cite to insurance executives’ unanimous testimony that their networks had to include at least one of the merging hospitals “to offer a product marketable to employers.” Id. Citing “strong, not equivocal” evidence, the court looked past its concerns that the insurers’ testimony could be “self-serving” and concluded that “the record as a whole support[ed] the view that insurers genuinely believe[d] that a plan that exclude[d] Advocate and North Shore [was] not viable in the North Shore Area.” Id. at *4–5. Here too, the Government realizes that testimony from this region’s major
insurers is an indispensable component of Dr. Smith’s analysis, if his proposed geographic markets are to correspond with the commercial realities of the southeastern Pennsylvania healthcare industry. See (Oct. 26, 2020 Oral Arg. Tr. 16:15–17:12, 238:4– 17, 316:21–317:13). Dr. Smith relies on insurer testimony for his conclusion that if the merger results in price increases for insurers, “they’d have to pass them onto their customers.” (Sept. 16, 2020 Hr’g Tr. (Smith) 52:16–22.) But as explained below, the insurers’ testimony here is not unanimous or unequivocal, is undercut by other evidence in the record and is ultimately not credible. Dr. Smith’s reliance, at face value, on what some of the insurers had to say is misplaced and unsupported by the record as a whole. Moreover, other than merely adopt the insurer testimony, Dr. Smith acknowledged that he had not done any other analysis to conclude that there is a correlation between patients and insurers, explaining only that it was “pretty basic economics.” (Id. at 52:23–24.) He testified
that he did not know “very many economists who would say that . . . when costs go up to a firm that’s competing in an imperfectly competitive market that there wouldn’t be some pass through.” (Id. at 52:24–53:2.) As a matter of academic econometric analysis, Dr. Smith could be correct, but relying on that simple principle is insufficient. Dr. Smith’s basic economics have to be supported by credible evidence that the insurers would have to agree to price increases instead of looking outside his proposed geographic markets. C The insurers’ views on this merger do not fill the gaps in Dr. Smith’s analysis. Such a conclusion is unavoidable when comparing the facts of this case to those in the
Third Circuit’s decision in Penn State Hershey, which concerned the proposed merger of the two largest hospitals in the Harrisburg area, Penn State Hershey Medical Center and Pinnacle Health System. See Penn State Hershey, 383 F.3d at 333. Hershey, the region’s leading academic medical center and provider of complex medical services, operated central Pennsylvania’s only specialty children’s hospital and the only heart- transplant center in the state outside of Philadelphia and Pittsburgh. See id. at 334; FTC v. Penn State Hershey Med. Ctr., 185 F. Supp. 3d 552, 554 (M.D. Pa. 2016), rev’d on other grounds, FTC v. Penn State Hershey Med. Ctr., 838 F.3d 327 (3d Cir. 2016). The FTC and Pennsylvania Attorney General sought to enjoin the merger, alleging that it would substantially lessen competition for GAC services in the relevant geographic market, defined as the four counties encompassing and immediately surrounding Harrisburg. See Penn State Hershey, 383 F.3d at 334, 338. The evidence showed that if allowed to merge, Hershey and Pinnacle would control seventy-six percent of GAC
services in that market. See id. at 334. The district court denied the government’s motion for a preliminary injunction, holding the government did not properly define the relevant geographic market. See id. at 339. The court based its decision largely on patient behavior, namely the number and percentage of Hershey’s patients who travel to Hershey from outside the four- county area. See id. at 339. The Third Circuit reversed, stating that the healthcare industry’s commercial realities dictated consideration of a possible price increase on insurers, not patients. See id. at 341–43. Crucial to the Circuit’s ruling was that “[t]he [g]overnment presented extensive evidence showing that insurers would have no choice but to accept a price increase from a combined Hershey/Pinnacle in lieu of excluding the
Hospitals from their networks.” Id. at 345 (emphasis added). The extensive evidence included testimony by two of central Pennsylvania’s largest insurers that they could not market a network to employers without at least one of the merging hospitals. See id. One of them estimated that it would lose half of its membership in a particular county if it excluded both hospitals from its network. See id. That insurer also stated it previously used the possibility of creating a network that included Hershey and another hospital in the Harrisburg market to get Pinnacle to accept lower prices. See id. Other insurers also used the merging hospitals’ independence to negotiate lower rate increases. See id. Moreover, the evidence showed that no health insurance product lacking access to at least one of the merging networks had been successfully marketed to employers in the relevant area. See id. The district court heard evidence of a “natural experiment” where an insurer excluded both of the merging hospitals from its health plan and saw
half its members switch to other health plans even though it priced its products much lower than its competitors. See id. at 345–46. That insurer was then told by brokers its network was no longer viable. See id. The evidence also included area employers’ testimony that it would be difficult to market a health plan without the merging hospitals. See id. at 345; see also Advocate Health Care, 841 F.3d at 465 (executives from several insurers testified unequivocally that their networks had to include one of the merging entities to be marketable); FTC v. Sanford Health, 926 F.3d 959, 963–64 (8th Cir. 2019) (representatives from state’s three largest insurance companies unequivocally testified their networks must include services controlled almost exclusively by the merging entities to be competitive). Finally, multiple payors testified
that they considered the proposed geographic market to be distinct and that they did not consider hospitals in other areas to be suitable alternatives. See Penn State Hershey, 383 F.3d at 346. The southeastern Pennsylvania and central Pennsylvania healthcare markets are vastly different. The district court in Penn State Hershey pointed to the nineteen hospitals within a sixty-five-mile drive of Harrisburg as suitable alternatives to Hershey and Pinnacle. See Penn State Hershey, 185 F. Supp. 3d at 557. Here, there are numerous health systems and many more hospitals within a far smaller radius. See supra Section I.B.iii. Those hospitals include highly regarded academic medical centers. See (Sept. 29, 2020 Hr’g Tr. (Klasko) 19:17–20:2 (Philadelphia area has a “monopolistic” health insurance market and “some of the greatest [provider] competition,” which includes six academic medical centers, giving the market some of the greatest concentration of hospitals in the country)); see also (DX0210-008 (showing
that Philadelphia has low hospital consolidation)); (Sept. 14, 2020 Hr’g Tr. (Markowitz) 65:10–66:2 (“when you compare it to other markets—[Philadelphia] is a less consolidated market than the other markets we’ve seen across the country”)). And again, while the region’s hospital consolidation is low, healthcare insurer consolidation is high. See supra Section I.C. Insurers in this region have numerous reputable providers they can include in their networks while some providers depend on being in the few major insurer networks to remain viable. See (Sept. 29, 2020 Hr’g Tr. (Klasko) 19:17–20:2, 29:5–10 (“traditionally, I couldn’t say no to any commercial insurer because . . . since we’re a safety net hospital and you’re so underpaid by the government, you’re actually losing
money on government sponsored services, that if you lost those commercial patients you really wouldn’t be able to survive very well.”)); cf. (Sept. 30, 2020 Hr’g Tr. (McTiernan) 57:6–11 (“to walk away from a commercial insurer is typically not a wise choice for health systems or hospitals to make . . . . I can’t really recall in Philadelphia that decision being made for a sustained period of time by any hospital or health system.”)). Most importantly, for the reasons that follow the Government here has not put forth anything like the “extensive evidence” discussed in Penn State Hershey proving that insurers in this region would accept a price increase in either of Dr. Smith’s proposed geographic markets. i Testimony from the region’s insurers on the Jefferson-Einstein merger’s potential competitive impact is neither unequivocal nor unanimous, something that reflects the competitive provider dynamic in southeastern Pennsylvania. The
Government does not rely at all on testimony or evidence from Aetna, the region’s second largest insurer, for good reason: Aetna has no concerns about the merger. See (PX7010, Investigational Hr’g Tr. (Morris) 93:15–21); see also (JX0062, Morris Dep. Tr. 31:13–32:16). Similarly, the Government did not rely on the area’s third biggest insurer, United, during the evidentiary hearing, oral argument or in any of its briefing. Cynthia Winings, United’s Vice President of Network Management, was not called to testify at the hearing. In a declaration and during her deposition, she expressed some concerns about the merger, including that she believed a network excluding the Jefferson and Einstein systems entirely “would not be marketable to residents of Montgomery or
Philadelphia Counties” and that the merger would “provide the additional leverage for Jefferson to ask for higher rates.” See (JX0064, Winings Dep. Tr. 329:14–331:5, 353:2– 354:10); (PX5007, Winings Decl. ¶¶ 19, 25).10 But she never said what the Government needed her to say—that the absence of
10 Many witnesses, in addition to being deposed and, for some, also testifying at the evidentiary hearing, submitted sworn declarations during the Government’s investigation. In certain circumstances, the witnesses’ declarations were less equivocal than their testimony. Defendants point out that the declarations were actually drafted by the FTC, while the FTC responds that it ensured that the statements in the declarations are “truthful, accurate, and clear” and “[t]here is no argument here that the witnesses attested to anything other than what they believed to be true.” See (Defs.’ Mot. in Limine at 3–4, ECF 132); (Pls.’ Resp. to Defs.’ Mot. in Limine at 4, ECF 166). The Court considers the declarations, but places greater weight on witnesses’ deposition or hearing testimony, where they were subject to cross-examination. the Jefferson and Einstein hospitals in the Government’s narrower Northern Philadelphia and Montgomery Areas would make United’s network unmarketable or that United would agree to pay higher rates for GAC services in the Government’s proposed markets. See generally (JX0064); (PX5007). In fact, she never said that
United would agree to pay higher rates to a combined Jefferson-Einstein anywhere. See generally (JX0064); (PX5007). And she never recognized the existence of distinct GAC markets in the Northern Philadelphia or Montgomery Areas. See generally (JX0064); (PX5007). Winings also acknowledged a United analysis demonstrating the company believes a broad range of alternative providers to EMCP or EMCM are available, including hospitals well outside of the Government’s Northern Philadelphia and Montgomery Areas such as Pottstown Hospital, HUP, Pennsylvania Hospital, Penn Presbyterian Medical Center, Main Line’s Lankenau Hospital and Methodist Hospital. See (JX0064, Winings Dep. Tr. 101:12–107:5); (DX0402-005–006).
ii Without Aetna or United, the Government relies principally on IBC and Cigna to support Dr. Smith’s algorithm and show that the geographic markets Smith has designed correspond to the commercial realities of this region’s healthcare industry. See (Oct. 26, 2020 Oral Arg. Tr. 16:15–18, 17:8–1211, 243:22–244:4). In fact, the Government maintains that the best record evidence supporting its position that the merged entity’s bargaining leverage will improve in its future contract negotiations with insurers is the combination of Dr. Smith’s “econometric analysis” and the
11 The oral argument transcript incorrectly names defense counsel as the speaker at 16:15–21, 17:2–7, 17:10 and 17:12. The Government provided each of these responses. testimony of IBC and Cigna representatives. See (id. at 238:4–17); see also (id. at 316:23–25, 317:1–5 (the best evidence that a hypothetical monopolist could impose a SSNIP in the Government’s defined geographic markets is “the testimony of the insurers again and the data analysis done by Dr. Smith.”)).
a The Merger Guidelines caution that while “[i]nformation from customers about how they would likely respond to a price increase, and the relative attractiveness of different products or suppliers, may be highly relevant, especially when corroborated by other evidence,” in evaluating customer testimony it is important to be “mindful that customers may oppose, or favor, a merger for reasons unrelated to the antitrust issues raised by that merger.” Merger Guidelines § 2.2.2. Courts in healthcare merger cases have expressed similar skepticism of insurer testimony and its potentially self-serving nature. See, e.g., FTC v. Tenet Health Care Corp., 186 F.3d 1045, 1054 (8th Cir. 1999) (reversing district court order enjoining a hospital merger, in part because of the
district court’s “reliance on the testimony of managed care payers, in the face of contrary evidence, that these for-profit entities would unhesitatingly accept a price increase rather than steer their subscribers to [other hospitals]. Without necessarily being disingenuous or self-serving or both, the testimony is at least contrary to the payers’ economic interests and thus is suspect.”); Advocate Health Care, 2017 WL 1022015, at *5 (“The Court shares some of defendants’ concerns about the credibility of the insurers’ testimony, which may indeed be self-serving . . . .”). Courts should not take at face value the testimony of insurers in hospital merger cases. Payor testimony must be corroborated by the record evidence taken as a whole. Above all, the testimony must be credible. It must be reasonable when considered in light of the other evidence and the insurers’ possible bias or motive must be assessed when determining how much weight the testimony should be given. The IBC and Cigna testimony the Government claims supports its markets is
not corroborated by the record evidence. And while Dr. Smith cites to portions of IBC and Cigna deposition testimony to bolster his rebuttal report, see, e.g., (Smith Rebuttal Rpt. ¶¶ 61 n.137, 88 n.188, 90 n.194–95, 183 n.326, 241 n.460–61), other portions of that testimony undercut Dr. Smith’s conclusions. b Paul Staudenmeier, IBC’s Vice President of Provider Contracting, Reimbursement and Value-Based Programs, testified to IBC’s alleged concerns with the merger, purportedly to show that if Jefferson and Einstein were allowed to combine, IBC would pay higher reimbursement rates if the merged entity demanded them. His testimony, considered against other record evidence and IBC’s motive to oppose the
merger for other reasons, was less than convincing. Staudenmeier’s concerns fluctuated in magnitude depending on how he was asked to give his view, and his trepidations were far less compelling when stated in an adversary setting versus an FTC declaration. At the evidentiary hearing, Staudenmeier testified that if after the merger Jefferson and Einstein demanded higher rates at their GAC hospitals, he “would expect it to be a difficult negotiation, and I would expect we would have to ultimately pay more if we wanted them to be in the network.” (Sept. 14, 2020 Hr’g Tr. (Staudenmeier 89:15– 90:10.) If Jefferson and Einstein’s GAC hospitals were not in IBC’s network, IBC “could probably from a technical standpoint sell a health plan, but we would be much less attractive to citizens who lived in” northern Philadelphia or Montgomery County. (Id. at 87:22–88:5.) Staudenmeier was most concerned about potential reductions in member choice
for certain GAC services in the Montgomery and Northern Philadelphia Areas if Jefferson and Einstein go out of IBC’s network after the transaction. He stated that if EMCP was not in IBC’s network, “[i]t would be very important to have Abington in the network” to serve members in northern Philadelphia and southern Montgomery County. (Id. at 86:9–23.) He also testified that to serve members in the geographic area that EMCM serves, if EMCM was not in-network for IBC then IBC “would need contracts with the Jefferson Lansdale facility, for example, or the Abington facility, for example. Not limited to those two, but they would be very important.” (Id. at 85:4–14.) Obstetric services would be particularly lacking in Montgomery County without EMCM or Abington Hospital. See (Id. at 85:20–25, 90:11–13); (JX0070, Staudenmeier Dep. Tr.
189:9–190:12); (PX5008, Staudenmeier Decl. ¶ 33). IBC’s opposition to this transaction exemplifies why courts should scrutinize insurer testimony in hospital merger cases. First of all, Staudenmeier believes that all hospital mergers are bad for consumers. See (Sept. 14, 2020 Hr’g Tr. (Staudenmeier) 123:20–124:23). He also says that IBC seeks to market networks that offer members “complete choice” and the “largest, broadest network of participating providers in the marketplace” such that “relative to competitors our goal is to have the largest network.” (Id. at 81:11–20.) This suggests that what IBC really wants to avoid is not having a larger network than its competitors, i.e., becoming any less dominant in the marketplace than it already is. But that “goal” has nothing to do with the actual issue the Government must prove—that IBC or other insurers would accept a SSNIP in the Government’s proposed geographic markets instead of turning to other providers in their networks outside those markets.
The record does not support Staudenmeier’s conclusory assertion that IBC would surrender to a price increase rather than steer its members to hospitals outside the Government’s markets. For instance, IBC threatened to terminate Einstein from its network before reaching its most recent agreement with Einstein and did not express a concern for network attractiveness or marketability in assessing Einstein’s termination. See (Sept. 16, 2020 Hr’g Tr. (Freedman (Einstein CEO)) 131:13–25); (DX0329-008 (stating IBC has sufficient network access/adequacy for commercial patients without Einstein and listing transfer of maternity cases, “negative press,” and “vulnerable hospital” among “key considerations/risks” relating to an Einstein termination)). IBC was also prepared to sever ties with Jefferson, including the then-recently acquired
Abington and Aria health systems, if its last negotiations failed, to a point where IBC’s CEO texted Jefferson’s CEO, Dr. Stephen Klasko, “basically a ‘take it or leave it’” offer. See (JX0070, Staudenmeier Dep. Tr. 271:8–13 (IBC’s message to Jefferson was that it “was prepared to move on if negotiations failed”)); (PX1375-004); (Sept. 14, 2020 Hr’g Tr. (DeAngelis) 298:23–299:19). IBC told Jefferson that given the breadth of provider competition in southeastern Pennsylvania, IBC’s network is not dependent on particular providers. See (Sept. 29, 2020 Hr’g Tr. (Klasko) 27:3–7 (“[W]hen it comes to commercial insurance, there’s so much provider competition that frankly anybody can, especially the big ones, United, Aetna, IBC, of course, and Cigna, they could just say fine, we won’t do it. And you don’t even have to ask me. You can ask IBC. IBC has told me that.”)). And further to Staudenmeier’s stated goal, IBC contracts with every health system in southeastern Pennsylvania, giving its subscribers numerous quality healthcare options. See (Sept.
14, 2020 Hr’g Tr. (Staudenmeier) 105:4–6). “Alternative providers offering a similar scope and quality of services in a similar geographic area limit the bargaining leverage of a provider and constrain its ability to negotiate higher reimbursement rates.” (Pls.’ FF ¶ 8.) Staudenmeier based his testimony about GAC hospital alternatives to Jefferson and Einstein on his purported “knowledge of the marketplace,” not any IBC analyses. See (JX0070, Staudenmeier Dep. Tr. 68:17–69:19, 192:19–194:3). But IBC’s own evaluation of suitable alternatives to Jefferson hospitals does not align with Staudenmeier’s belief. When IBC considered terminating Abington, Abington- Lansdale, Methodist, and TJUH from its network, it identified HUP, Pennsylvania
Hospital, St. Mary Medical Center and Holy Redeemer, not Einstein, as redirection hospitals. See (DX0323-011); see also (JX0070, Staudenmeier Dep. Tr. 188:6–11). Additionally, Staudenmeier singled out EMCM, Abington and Abington-Lansdale’s provision of obstetric services as a main reason why IBC’s network would suffer without those facilities in it, but his concern lacks a foundation. Staudenmeier said that the volume of IBC member deliveries could not be taken on by all other providers in IBC’s network, but that was a “gut reaction.” See (JX0070, Staudenmeier Dep. Tr. 307:5–9). He did not know how many IBC members residing in Montgomery County deliver at EMCM, Abington or Abington-Lansdale. See (id. at 190:13–191:20). But he acknowledged that IBC contracts with Holy Redeemer, which offers obstetric services in Montgomery County and is located barely outside the Government’s Montgomery Area market. See (id. at 318:13–18). He also recognized that HUP is a respected obstetric services provider but did not know what percentage of IBC members residing in
Montgomery County receive obstetrics care at Penn. See (id. at 195:6–196:2).12 The record evidence also shows that IBC has a clear motive, other than antitrust concerns, to oppose this merger. As Staudenmeier confirmed, IBC considers Jefferson to be a potential competitive threat as an insurer. See (Sept. 14, 2020 Hr’g Tr. (Staudenmeier) 113:16–114:8). IBC recognizes, and does not like, that “the merger would take Jefferson from being less of a potential competitor to IBC [to] more of an actual competitor.” See (id. at 116:23–117:1). Specifically, Jefferson and Einstein currently compete with IBC as insurers through their ownership in Health Partners Plans, Inc. (“HPP”), a Medicaid and Medicare insurer that covers approximately 256,000 lives in southeastern Pennsylvania. See (id. at 116:10–13); (DX9601); (JX0071,
Snyder Dep. Tr. 148:1–4). IBC has the largest Medicaid plan in Philadelphia and HPP is the second largest. See (Sept. 16, 2020 Hr’g Tr. (Freedman) 114:13–20). Jefferson and Einstein each own twenty-five percent of HPP, and a combined Jefferson-Einstein would thus own half the entity. See (DX9601). Temple University owns the remaining
12 In any event, the relevant product market is GAC services, not obstetric services in isolation. In his Government-drafted declaration, Staudenmeier states that “[a] provider with a favorable bargaining position regarding one service line can use that bargaining position in negotiations with Independence to extract higher rates overall, including rates relating to other service lines. For example, a provider that has a favorable bargaining position relating to obstetrics services . . . could use that bargaining position to negotiate for higher rates overall, including rates for services other than obstetrics.” (PX5008, Staudenmeier Decl. ¶ 11.) But Staudenmeier’s opinion, based as it is on his “gut reaction,” does not show that if a combined Jefferson-Einstein could have a more favorable position for obstetric services, IBC, or any other insurer, would then have to pay the merged entity higher rates for the overall cluster of GAC services. fifty percent of HPP and, through a separate transaction, Jefferson is attempting to purchase Temple’s share. See (id.). IBC’s CEO reacted to Jefferson’s potentially growing position in HPP by saying “it remains to be seen how we are going to be able to collaborate with anyone who is in direct competition with us as an insurer.” (DX9601-
003.) IBC does not want healthcare providers competing with it in the health insurance market. See (Sept. 30, 2020 Hr’g Tr. (McTiernan) 56:4–12 (based on his experience working at IBC, IBC “viewed hospitals wanting to get into the insurance business, no matter who the hospital or health system was, as a potential competitive threat . . . .”)). For example, before Einstein and Jefferson agreed to merge, Einstein discussed a potential partnership with a Pittsburgh-based provider and insurer, the University of Pittsburgh Medical Center (“UPMC”). See (Sept. 16, 2020 Hr’g Tr. (Freedman) 132:22–134:2); (id. (Maksimow (SVP Partnerships, Mergers and Acquisitions, Kaufman Hall)) at 224:19–225:8). As part of a potential UPMC/Einstein
deal, UPMC wanted to bring Einstein together with a managed care company so that UPMC could enter the southeastern Pennsylvania market as both a provider and an insurer. See (id. (Freedman) at 132:22–134:2); (id. (Maksimow) at 224:19–225:8). While IBC and Einstein were negotiating a new contract, IBC invited Einstein’s President and CEO, Barry Freedman, to an Eagles game. See (id. (Freedman) at 131:13–25). At halftime, IBC’s CEO and another executive told Freedman that IBC would terminate its contract with Einstein if Einstein partnered with UPMC. See (id.); (Sept. 30, 2020 Hr’g Tr. (DeAngelis) 53:2–18). Shortly thereafter, IBC sent Einstein a contract proposal containing significant rate reductions in comparison to a prior proposal. See (Sept. 30, 2020 Hr’g Tr. (DeAngelis) 53:21–22). Einstein termed it the “UPMC proposal” because “the rate reductions were so significant that it was not founded on any basis or rationale.” (Id. at 53:22–54:9.) Ultimately, IBC’s resistance to UPMC played a role in Einstein and UPMC’s inability to reach a final agreement. See
(Sept. 16, 2020 Hr’g Tr. (Freedman) 133:13–20). Also, before Jefferson and Einstein agreed to merge, IBC’s parent company IHG grew concerned that Jefferson and Main Line Health “could become direct competitors to IHG (ala UPMC in Western PA).” (DX332-005.) IBC’s CEO called Dr. Klasko and told him Jefferson “need[ed] to not be in [an Accountable Care Organization] with Main Line Health because they were going to go against Main Line Health, and it would be much better for Jefferson if I don’t do anything with [the CEO of Main Line].” (Sept. 29, 2020 Hr’g Tr. (Klasko) 58:2–8.) Notably, IHG contemplated responding to competition from Jefferson and Main Line by excluding both systems from its network entirely and building tiered network
products around other providers, such as Penn Medicine. See (DX332-018–019); (Sept. 14, 2020 Hr’g Tr. (Staudenmeier) 115:6–18). IHG also considered employing a strategy of “[f]eatur[ing] member costs sharing models that . . . disincentivize use of Jefferson [and] Main Line Health providers.” See (DX0332-023); (Sept. 14, 2020 Hr’g Tr. (Staudenmeier) 115:23–116:4). By this time, Abington and Abington-Lansdale were already part of the Jefferson Health System and Paoli and Bryn Mawr, two hospitals also in the Government’s Montgomery Area GAC market, were part of the Main Line Health System. IHG was not worried about network marketability when it considered excluding, or disincentivizing members from using, all of Jefferson and Main Line Health’s hospitals. See generally (DX0332). IBC’s contention that it needs a combined Jefferson-Einstein in its network so much that it would pay higher rates is called into question by its contemplation of a network without any of Jefferson or Main Line’s hospitals.
IBC’s views on this merger merit little weight. The evidence does not support the assertion that IBC would roll over and pay higher prices instead of turning to its network of hospitals outside the Government’s proposed geographic markets. What the record does show is that IBC is far more concerned with hospitals joining forces where, as here, IBC views them as a competitive threat in the insurance market. c Cigna’s perspective on the merger does not help the Government meet its burden of proof either. Keith Markowitz, Cigna’s Regional Director of Operations and Marketing, testified that Cigna, the smallest of the region’s major insurers, could not offer a marketable health plan to its Montgomery County members that excluded
EMCM, Abington and Abington-Lansdale or to members “around the Einstein Philadelphia Hospital” that excluded EMCP and Abington. See (Sept. 14, 2020 Hr’g Tr. (Markowitz) 23:13–21, 28:1–8). Markowitz thought the only good alternatives to EMCM are Abington and Abington-Lansdale. See (id. at 22:7–23:8). He also testified there were no alternatives to EMCP and Abington “with the same acuity of service” in the EMCP area. See (id. at 28:9–12). He therefore concluded that Cigna would pay higher rates to include EMCM, Abington, Abington Lansdale and EMCP in its network. See (id. at 26:18–27:1, 29:9–16). On balance, the evidence does not support such a conclusion. First, Markowitz overstated Jefferson and Einstein’s significance to Cigna’s networks. Cigna believes no one health provider in southeastern Pennsylvania has a substantial part of the market. See, e.g., (JX0066, Markowitz Dep. Tr. 179:17–180:5); (DX0210-008). Its network does not rely on the Jefferson system, and there is nothing
to indicate that Cigna will rely on Jefferson once Jefferson acquires a health system with an approximately twenty percent commercial payer mix. Second, Markowitz does not think Einstein provides as high acuity services as Jefferson. See (JX0066, Markowitz Dep. Tr. 341:6–22). Rather, he recognized that the Jefferson, Penn and Temple health systems provide the highest acuity of care and that Jefferson and Penn are each other’s largest competitors. See (id. at 338:3–339:8); (Sept. 14, 2020 Hr’g Tr. (Markowitz) 63:7–10, 67:13–15). Markowitz stated a number of times that acuity of services factored into his assessment of good alternatives to EMCM and EMCP. See, e.g., (Sept. 14, 2020 Hr’g Tr. (Markowitz) 22:14–21, 27:13–25, 28:1–12). But he did not explain why he believed EMCM and EMCP’s only good alternatives are
higher acuity facilities than EMCM and EMCP themselves. Third, although he spoke about them when prompted by the Government, Markowitz did not appear to conceive of the Northern Philadelphia and Montgomery Areas as submarkets that Cigna considers when building a network. Cigna’s commercial population predominantly spans the four counties surrounding the City of Philadelphia and Markowitz evaluates Cigna’s network marketability within a broader geography than the Northern Philadelphia and Montgomery Areas, both of which include only pieces of two counties in southeastern Pennsylvania. See (JX0066, Markowitz Dep. Tr. 93:11–17 (“[I]n my opinion, and I’ve been doing this for a long time, it is with very rare circumstances that you have a customer, a client, an employer group, that is contained within one geographic county, and that—so that a network that is within a county geography is not marketable to a broad number of employer groups.”)); see also (id. at 310:2–15 (for commercial insurance, “you couldn’t sell a
product if you were just in Philadelphia and Bucks County because invariably people, if they’re located in Philadelphia, would live within the entire metropolitan area.”)). Markowitz said there is a submarket around EMCM “predominantly in Montgomery County,” but did not say his definition of the submarket aligns with the Government’s Montgomery Area. See (id. at 334:1–6). Moreover, Markowitz could not identify a submarket for GAC services around EMCP until prompted to do so, twice, by the Government at his deposition. At first, when asked whether he believes there is a submarket around EMCP, Markowitz said people in the EMCP area would want to go to EMCP but “when you get into the Philadelphia area, I think that becomes a little bit more difficult, because when you’re close in the city, the hospitals are much closer
together.” See (id. at 334:7–19). It was not until counsel for the Government changed his question to, “is the submarket around [EMCP] primarily in the North Philadelphia area?” that Markowitz agreed a submarket exists around EMCP in North Philadelphia pushing into Montgomery County. See (id. at 334:20–335:7). Even then, he did not identify more specific boundaries for the submarket. See generally (id.). It makes little sense that Cigna would pay higher reimbursement rates to hospitals because of their purported importance to areas Cigna does not believe to be distinct submarkets. iii The “extensive evidence” that figured so prominently in Penn State Hershey showing that insurers in central Pennsylvania would have no choice but to accept a price increase does not exist in this case. First and foremost, the second largest health insurer in southeastern Pennsylvania has “no concerns” about the Jefferson-Einstein merger and the third largest never said it would pay higher rates for GAC services
post-merger. Given the numerous healthcare systems here, no insurer can credibly assert that there would be “no network” without a combined Jefferson and Einstein— something the insurers could say when Hershey and Pinnacle, the two largest Harrisburg area hospitals (which together would have controlled seventy-six percent of GAC services in that market), attempted to merge. The Court lacks the benefit, and evidence, of the results of any “natural experiment” where an insurer tried to market a plan without Jefferson and Einstein in it and lost half of its membership as a result—despite its plan being much cheaper than its competitors. And no employers testified they would have a difficult time marketing a health plan without Jefferson and Einstein. To the contrary, the one employer who
did weigh in, a large school district, said its employees would be fine with a health plan excluding the two systems. See (JX0051, Demkin (Benefit Specialist, Lower Merion Sch. Dist.) Dep. Tr. 15:9–17, 18:15–19:12, 62:5–19, 62:23–63:4); see also (id. at 50:11–25 (employees have many options for inpatient GAC services and, if a facility ever became too costly, employees could go to another provider down the street)). The Government has not proven that the Government’s Northern Philadelphia and Montgomery Areas correspond to the commercial realities of the southeastern Pennsylvania healthcare industry.13
13 The Government attempts to make much of evidence in the record showing that Jefferson and Einstein hope to make themselves more indispensable to insurers through this merger. See, e.g., D Nor does the Government’s innovation of a third market for inpatient Acute Rehabilitation Services provide a basis for enjoining the merger. The Hospitals do not agree that the Government has properly defined the relevant product market to include
only services provided at IRFs, and not a cluster of the same services regardless of whether provided at an IRF or in another post-acute care setting like a high-end SNF. (Defs.’ Proposed Conclusions of Law ¶¶ 9–10.) Even assuming the relevant product market appropriately includes only those services provided at IRFs, the Government still fails to meet its burden to establish a relevant geographic market for inpatient Acute Rehabilitation Services. As it did for its GAC geographic markets, the Government relies on econometrics and insurer testimony to prove the propriety of its proposed Philadelphia Area market. But it has not shown that the market corresponds with commercial realities and it thus cannot pass the HMT. i
To define the relevant geographic market for Acute Rehabilitation Services, Dr. Smith “follow[ed] the same methodology” he used to define the GAC services geographic markets. (Smith Rpt. ¶ 147.) First, he determined “that MossRehab at Elkins Park’s closest Jefferson substitute is Magee . . . with a diversion ratio of 18.5 [percent] from MossRehab at Elkins Park to Magee.” (Id.) Then, because “no third-party IRF ha[s] a higher diversion ratio,” he did not add any additional IRFs. See (id.). He then added to
(Pls.’ FF ¶ 57); (JX0022-012); (PX6041-003); (PX6066-047). This misses the point. The Hospitals’ aspirations do not change the fact that this record does not show, through the lens of the insurers, that insurers would pay a price increase for GAC services in the Northern Philadelphia and Montgomery Areas instead of looking to hospitals outside those markets. the candidate market “all IRFs at least as close to MossRehab at Elkins Park as Magee in terms of drive distance, as well as Penn Rehab” (i.e., Good Shepherd Penn Partners) because it “is less than one-mile farther from MossRehab at Elkins Park than Magee.” (Id.) Smith then tested whether his candidate market satisfied the HMT “by
calculating the level of intra-market substitution from MossRehab at Elkins Park.”14 (Id.) Concluding that it did, he defined the Philadelphia Area as a “relevant antitrust market.” See (id. at ¶ 148, Fig. 5). Here again, Dr. Smith assumed correlated behavior between patients and insurers in the market for Acute Rehabilitation Services when he uses patient diversion ratios to select a candidate market. But “patients’ response to the price increase demanded by the post-Merger” health systems “is not important.” Penn State Hershey, 838 F.3d at 343. What matters is how insurers will behave. Dr. Smith’s reliance on patient diversion ratios to construct a candidate market for Acute Rehabilitation Services is appropriate only if there is evidence of correlation between insurer and
patient behavior regarding provider substitution. Even then, Dr. Smith’s diversion ratios imperfectly capture all the factors underlying patient “choices” and insurer responses in the rehabilitation context. Patient placements can depend on factors like the location of available beds at the time of discharge to rehabilitation care. See (Sept. 29, 2020 Hr’g Tr. (Seminara) 221:16–23 (“the first one that can take the patient is the one that's going to get the referral”)). Rehabilitation placement preferences may vary depending on underlying injuries (e.g.,
14 According to Dr. Smith, the “aggregate diversion ratio from MossRehab at Elkins Park to this set of IRFs is [fifty percent], over twice the [twenty percent] aggregate diversion ratio necessary for a hypothetical monopolist to profitably increase prices by a [five percent] SSNIP at MossRehab at Elkins Park in negotiations with commercial insurers . . . .” (Smith Rpt. ¶ 148.) spinal cord injury versus stroke). See (Sept. 14, 2020 Hr’g Tr. (Staudenmeier) 97:20–22 (acknowledging patients would prefer Magee or Moss “for traumatic brain and spinal cord injuries”)). And, importantly, patient volume is small. See (Oct. 1, 2020 Hr’g Tr. (Smith) 28:11–13, ECF No. 263 (“[O]ne feature of inpatient acute rehabilitation services
versus GAC inpatient services for example is there’s many fewer discharges.”)). Diversion ratios are only “statistical estimates.” (Id. at 68:23–69:1.) Nevertheless, Dr. Smith’s market construction does not explicitly consider a range of error associated with his estimates. Instead, when using diversion ratios to select a candidate set of facilities for a relevant market” he “use[d] the point estimate first” and then tested “whether that market satisfies the [HMT.]” (Id. at 69:9–16.) Dr. Smith’s analysis does not account for the possibility that there are “different ways of estimating diversions” that may lead to different results. (Sept. 29, 2020 Hr’g Tr. (Ramanarayanan) 277:18–278:1.) And because the number of patients receiving inpatient Acute Rehabilitation Services is so small, “small” changes to diversion ratio
estimates can “change[ ] the contours of the geographic market pretty dramatically.” (Id. at 277:9–13); see also (Ramanarayanan Rpt. ¶ 137 (finding Dr. Smith overstated diversion ratios where, for example, he incorrectly classified Magee patients insured with automobile insurance as “commercially insured”)). The Government’s evidence, when viewed through the lens of the insurers, does not support Dr. Smith’s conclusion that patient diversion ratios will predict the anticompetitive consequences of a possible Jefferson/Einstein merger in Acute Rehabilitation Services. As “qualitative evidence” in support of his proposed Philadelphia Area geographic market, Dr. Smith again cited to insurer declarations “that a health plan network that excluded the Defendants’ in-market IRFs, but included all other IRFs in and around Philadelphia—including Bryn Mawr Rehab, St. Mary Rehab, and IRF units—would be difficult or impossible to market to residents of Philadelphia County or Montgomery County, or any geographic area that included
either of those counties.” (Smith Rpt. ¶ 150); see also (PX5008, Staudenmeier Decl. ¶ 32); (PX5007, Winings Decl. ¶ 22); (PX5006, Markowitz Decl. ¶ 23). The Government, as it did with inpatient GAC services, relies primarily on the testimony of IBC’s Staudenmeier and Cigna’s Markowitz. See (Oct. 26, 2020 Oral Arg. Tr. 243:17– 244:4). The Government’s lack of evidence that insurers would pay higher prices in the Government’s proposed Philadelphia Area is even more pronounced when it comes to Acute Rehabilitation Services. ii Rehabilitation services play a minor role in health systems’ operations and contracts. See (DDX010-035 (illustrating the share of Einstein and Jefferson’s
commercial revenue from inpatient rehabilitation)). Relatively few health plan members ever require inpatient acute rehabilitation services. In 2018, for example, only 0.01 percent of Jefferson and Einstein’s total commercial patient visits were for inpatient rehabilitation. See (id. (showing Jefferson and Einstein had 991 commercial inpatient rehabilitation patient visits out of approximately 800,000 total commercial patient visits)); cf. (Smith Rpt. Table 5 (estimating, based on annual average commercial discharges between 2016 and 2018, that Jefferson and Einstein have 826 annual commercial IRF discharges)). Inpatient rehabilitation accounted for just 2.6 percent of Jefferson and Einstein’s total commercial revenue. See (DDX010-035). Acute Rehabilitation Services are just not a focus in insurer-member contracting. United’s Winings has never heard of an employer inquiring about the rehabilitation hospitals included in her company’s southeastern Pennsylvania network. See (JX0064, Winings Dep. Tr. 309:17–310:2). Markowitz has not discussed rehabilitation facilities
with employers for at least five years. See (Sept. 14, 2020 Hr’g Tr. (Markowitz) 66:18– 67:5); (JX0066, Markowitz Dep. Tr. 290:5–12). Staudenmeier acknowledged “inpatient rehab services isn’t something you look at when you’re selecting your health plan.” (Sept. 14, 2020 Hr’g Tr. (Staudenmeier) 97:3–4.) The only employer to testify confirmed it does not select health plans based on the inpatient acute rehabilitation services providers they include or exclude. See (JX0051, Demkin Dep. Tr. 59:10–14); see also (id. at 59:23–60:5 (employees “could be well served by a health plan so long as it included some provider of acute rehab services.”) (emphasis added)). Because members do not commonly inquire about Acute Rehabilitation Services when choosing their health plans, insurers’ network marketability is seemingly
unaffected by the particular IRFs in-network. Cf. (Sept. 29, 2020 Hr’g Tr. (Ramanarayanan) 290:2–23 (“[H]ealth insurers have the ability . . . to selectively contract with a few inpatient rehab providers and offer a health plan network which would still be attractive to members and employers because they don’t care as much about the identity of specific providers in the network.”)). According to Markowitz, acute rehabilitation facilities only impact Cigna’s network marketability in that they are “a checkmark.” See (Sept. 14, 2020 Hr’g Tr. (Markowitz) 69:9–23). If that’s the case, insurers can contract with numerous reputable IRFs outside the Government’s proposed Philadelphia Area, including Bryn Mawr Rehab and St. Mary Rehab, to “check a box” and have IRFs in-network. See, e.g., (JX0066, Markowitz Dep. Tr. 217:10–15 (acknowledging that Cigna contracts with Bryn Mawr Rehab and St. Mary Rehab)); (Sept. 14, 2020 Hr’g Tr. (Markowitz) 40:1–4 (stating Bryn Mawr Rehab has “a significant reputation”)); cf. (DX8517-036 (“Philadelphia and Montgomery Counties are
highly saturated with post-acute care facilities . . . .”)); (Sept. 29, 2020 Hr’g Tr. (Seminara) 229:6–17 (describing the competitive landscape for inpatient rehabilitation services in southeastern Pennsylvania as “vicious” and saying that professionals in other areas “can’t believe that we have this competition and this competitive market”)). Providers similarly consider Acute Rehabilitation Services “less important than most other services” in insurer-provider negotiations. See (Sept. 30, 2020 Hr’g Tr. (McTiernan) 65:1–12.); cf. (Ramanarayanan Rpt. ¶ 180 n.396 (citing JX0025, Phillips (President, Bryn Mawr Rehab) Dep. Tr. 65:9–17 (Main Line Health’s primary focus in negotiations with commercial insurers is general acute care); JX0034, Buongiorno Dep. Tr. 95:19–96:6 (Main Line Health spends more time on acute care rates than
rehabilitation rates in commercial insurer negotiations); JX0063, Robinson (VP Managed Care Contracting, Trinity Mid-Atlantic) Dep. Tr. 36:19–37:4 (Trinity Mid- Atlantic’s focus on inpatient rehab rates is “of a much smaller scale” compared with general acute care rates in commercial insurer negotiations)). Acute Rehabilitation Services contribute marginally to insurer revenues from provider contracts. See (Sept. 30, 2020 Hr’g Tr. (McTiernan) 65:13–66:1 (“[F]rom the insurance perspective,” rehabilitation rates are “a very small percentage of their overall medical spend. So it was not necessarily on their radar as the most important thing.”); (DX9475-001 (Magee accounted for $1.5 million out of $121.8 million (approximately 1.2 percent) in enterprise base revenue associated with United-Jefferson 2018 contract negotiation)); (DX9475-004 (Magee accounted for $8.1 out of $504.4 million (approximately 1.6 percent) in enterprise base revenue associated with Aetna-Jefferson 2018 contract)). Against this backdrop, the Government expects the Court to believe that IBC
and Cigna could not offer a marketable health plan if their competitors’ networks included both Jefferson and Einstein IRFs, particularly Moss and Magee, but theirs did not. See (Sept. 14, 2020 Hr’g Tr. (Markowitz) 43:8–25); (id. (Staudenmeier) at 96:20– 97:1). Staudenmeier testified that if a merged Jefferson/Einstein demanded higher IRF reimbursement rates, “focused on unit cost and medical cost, ultimately if I was required, and I suspect our marketing folks would require me to come to agreement, I would have to pay more for those services that are due today.” See (id. (Staudenmeier) at 100:12–25). Markowitz testified that Cigna “would likely have to pay . . . increased rates” if a merged Jefferson/Einstein demanded them for their IRFs because it “would need to have a marketable network to our customers.” (Id. (Markowitz) at 44:1–8.)15
The record belies these conclusory assertions. To begin, Staudenmeier and Markowitz have no experience negotiating health plans that exclude Moss or Magee. See (Oct. 26, 2020 Oral Arg. Tr. 70:5–11). There are no documents showing IBC or Cigna believe their networks would be unmarketable to southeastern Pennsylvania members without Moss, Magee or any IRF in the Government’s Philadelphia Area. And there is none of the other evidence relevant to the Penn State Hershey case
15 Consistent with its lack of concern for the merger, Aetna did not express a view on the merger’s potential effects on inpatient acute rehabilitation services. See generally (JX0062). Winings never said that United would agree to pay higher prices to keep Moss, Magee or any particular IRF(s) in its network. See generally (JX0064); (PX5007). showing insurers would have to pay higher prices for Acute Rehabilitation Service in the Philadelphia Area. The relative insignificance of inpatient rehabilitation services to members and insurers contradicts the insurers’ statements that they would accept price increases for
inpatient rehabilitation services. In fact, the services affect so little of their bottom line and so few of their members that they serve as only a checkmark in contract negotiations.16 Bare assertions that contradict insurers’ financial interests are insufficient proof of their genuine likely response to a price increase for Acute Rehabilitation Services in the Philadelphia Area. IBC’s testimony in particular is rank speculation. After admitting inpatient rehabilitation services are not a consideration when individuals or employers select their health plans, Staudenmeier nonetheless went on to say that having Jefferson and Einstein IRFs out of IBC’s network “would cause us a problem” with a patient who wished to go to one of those hospitals and “it would be a problem on renewal of that
group or individual contract” associated with the patient. (Sept. 14, 2020 Hr’g Tr. (Staudenmeier) 97:4–11.) In other words, he believes IBC would pay higher prices for Jefferson and Einstein IRFs because an insured, at some point during the life of the insurance contract, might decide she needed Moss or Magee. See (Oct. 26, 2020 Oral Arg. Tr. 249:1–250:6); see also (Sept. 14, 2020 Hr’g Tr. (Staudenmeier) 97:2–11). The ripple effect from this lone patient’s decision would then purportedly lead to a price
16 The Court is not suggesting that insurers would never accept a price increase from a hypothetical monopolist for a product that represents a small overall percentage of provider care. There may well exist a small cluster of, or even individual, healthcare services so significant to members and a geographic area that insurers would pay higher reimbursement rates to keep them in their networks. On the facts of this case, Acute Rehabilitation Services in the Philadelphia Area is not among them. increase at renewal time, notwithstanding that Acute Rehabilitation Services were little more than an afterthought when the existing contract was negotiated. See (Oct. 26, 2020 Oral Arg. Tr. 249:1–250:6); see also (Sept. 14, 2020 Hr’g Tr. (Staudenmeier) 97:2–11). The Government cannot satisfy its burden of proof relying on a theory like
that. Insurer testimony does not validate the Government’s Philadelphia Area. Having failed to establish a relevant geographic market for Acute Rehabilitation Services, the Government has not proven that Acute Rehabilitation Services sold to commercial insurers and their members in the Philadelphia Area is a relevant market within which the competitive effects of this merger can be properly assessed. E “[I]dentification of a proper market is a ‘necessary predicate’ to the [Government’s] task of” demonstrating a likelihood of success on the merits. FTC v. Freeman Hosp., 69 F.3d 260, 268 (8th Cir. 1995) (“Without a well-defined relevant
market, an examination of a transaction’s competitive effects is without context or meaning.”); see also Tenet Health Care, 186 F.3d at 1051 (“it is . . . essential that the [government] identify a credible relevant market before a preliminary injunction may properly issue”) (citation omitted); cf. Penn State Hershey, 838 F.3d at 337–38. Because the Government has not proven that the Northern Philadelphia and Montgomery Areas are relevant geographic markets for GAC services, or that the Philadelphia Area is a relevant geographic market for Acute Rehabilitation Services, the Government has not identified a single relevant market to make its prima facie case. See RAG-Stiftung, 436 F. Supp. 3d at 287 (“[B]ecause evaluating a merger’s competitive effects on a market requires the [government] to properly define a market in terms of both product and geography,” the government’s improper relevant product definition “all but precludes the Court from siding with it.”). The Government has not shown it is likely to succeed on the merits of its case at the administrative hearing.
IV Finally, the weighing of equities is statutorily required. 15 U.S.C. § 53(b) (“Upon a proper showing that, weighing the equities and considering the Commission's likelihood of ultimate success, such action would be in the public interest . . . a preliminary injunction may be granted . . . .”). Although Section 13(b) “is silent as to what specifically those equities are,” they require consideration of whether granting an injunction would be in the public’s interest. Penn State Hershey, 838 F.3d at 352. That is, whether the harm the Hospitals will suffer if the merger is delayed will harm the public more than if the injunction is not issued. Id. Where, as here, the Government has not shown it is likely to prevail on the merits of its Section 7 claim, the equities do
not favor its requested relief. “[T]he propriety of a preliminary injunction typically rises and falls with the FTC’s likelihood of success on the merits and the public interest in enforcing antitrust laws.” RAG-Stiftung, 436 F. Supp. 3d at 321 (concluding there was “no need for the Court to resolve . . . competing equities” where the FTC had not shown that it was “likely that the proposed merger will substantially lessen competition for a particular product in a particular geographic market”). The public’s interest in “effective enforcement” is “[t]he principal equity weighing in favor of” an injunction. Penn State Hershey, 838 F.3d at 352. It may be “extraordinarily difficult to unscramble the egg” after a merger is consummated if the FTC’s merits trial later determines the combination violates Section 7. Id. at 353 (citation omitted). But to allow an injunction on this basis alone would vitiate the Government’s obligation to show that it is likely to succeed on the merits of its claim that the proposed merger is anticompetitive. Cf. FTC
v. Great Lakes Chem. Corp., 528 F. Supp. 84, 99 (N.D. Ill. 1981) (“Section 13(b) was not enacted to authorize automatic preliminary injunctions . . . .”). The Government has not shown “that there is a credible threat of harm to competition during the time between the denial of this preliminary injunction and the final adjudication of” the merits because it has not established its prima facie case. FTC v. Freeman Hosp., 911 F. Supp. 1213, 1227 (W.D. Mo.), aff'd, 69 F.3d 260, 272 (8th Cir. 1995)(“[O]nce the district court found the FTC had failed to show a substantial threat to competitive hospital prices . . . the public equities advanced by the FTC were substantially weakened.”). Whatever weakened equities the Government could argue cannot justify enjoining this transaction given its failure to show a likelihood of success
on the merits. An appropriate Order follows. BY THE COURT:
_/s_/_ _G_e_r_a__ld__ J__. _P_a__p_p_e_r_t_ __ GERALD J. PAPPERT, J.
FEDERAL TRADE COMMISSION v. THOMAS JEFFERSON UNIVERSITY (FEDERAL TRADE COMMISSION v. THOMAS JEFFERSON UNIVERSITY) — published by Counsel Stack Legal Research, free access to 12M+ legal documents.